Text review methods, devices, electronic equipment, storage media and procedures
By conducting differentiated review of public and sensitive fields in text, and by using public internet channels and non-public intranet channels to process public and sensitive information respectively, the inefficiency, lag, and security issues of existing text review methods are resolved, achieving efficient and secure text review.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- INDUSTRIAL AND COMMERCIAL BANK OF CHINA
- Filing Date
- 2025-08-29
- Publication Date
- 2026-07-31
AI Technical Summary
Existing text review methods are inefficient, information verification is slow, review coverage is limited, and data security is difficult to guarantee.
By acquiring the target text to be reviewed, parsing its public and sensitive fields, and using matching review methods, the data is reviewed through both public internet channels and private intranet channels to ensure data security.
It improved the efficiency and coverage of text review, ensured the security of core data, and avoided the risk of information leakage.
Smart Images
Figure CN122489746A_ABST
Abstract
Description
Technical Field
[0001] The embodiments of the present invention relate to the field of artificial intelligence, specifically the application of large models in the fields of information security and fintech, and particularly to a text review method, apparatus, electronic device, storage medium, and program. Background Technology
[0002] In today's information explosion, data is being generated and flowing at an unprecedented speed and scale, with various types of information, both true and false, intertwined. This has not only dramatically increased the difficulty of compliance supervision, but also led to increasingly stringent requirements. However, existing manual text review methods have many drawbacks: low review efficiency, delayed information verification, and limited review coverage. Although some automated text review methods have emerged in existing technologies, these methods usually require transmitting text to external servers for processing, making it difficult to ensure the security of text data. Summary of the Invention
[0003] This invention provides a text review method, apparatus, electronic device, storage medium, and program that can improve the efficiency and coverage of text review while ensuring the security of core data.
[0004] According to one aspect of the present invention, a text review method is provided, comprising:
[0005] Obtain the target text to be reviewed;
[0006] The target text to be reviewed is parsed to determine the target reference fields of the target text to be reviewed; wherein, the target reference fields include target public fields and target sensitive fields;
[0007] The content of the target reference field is reviewed using a matching text review method based on the type of the target reference field.
[0008] According to another aspect of the present invention, a text review device is provided, comprising:
[0009] The target text to be reviewed acquisition module is used to acquire the target text to be reviewed;
[0010] The target reference field determination module is used to parse the target text to be reviewed and determine the target reference fields of the target text to be reviewed; wherein, the target reference fields include target public fields and target sensitive fields;
[0011] The text review module is used to review the content of the target reference field using a matching text review method based on the type of the target reference field.
[0012] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:
[0013] At least one processor; and
[0014] A memory communicatively connected to the at least one processor; wherein,
[0015] The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the text review method according to any embodiment of the present invention.
[0016] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the text review method described in any embodiment of the present invention.
[0017] According to another aspect of the present invention, a computer program product is also provided, comprising a computer program that, when executed by a processor, implements the text review method described in any embodiment of the present invention.
[0018] This invention, through obtaining and parsing the target text to be reviewed, determines the target reference fields. These target reference fields include publicly accessible fields and sensitive fields. After determining the publicly accessible and sensitive fields, the content of the target reference fields is reviewed using a matching text review method based on their type. This solution addresses the problems of low review efficiency, delayed information verification, and insufficient review coverage in existing text review methods, improving both the efficiency and coverage of text review while ensuring the security of core data.
[0019] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description
[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0021] Figure 1 This is a flowchart of a text review method provided in Embodiment 1 of the present invention;
[0022] Figure 2 This is a flowchart of a text review method provided in Embodiment 2 of the present invention;
[0023] Figure 3 This is a framework diagram of a text review system provided in Embodiment 2 of the present invention;
[0024] Figure 4 This is a schematic diagram of a text review device provided in Embodiment 3 of the present invention;
[0025] Figure 5 This is a schematic diagram of the structure of an electronic device provided in Embodiment 4 of the present invention. Detailed Implementation
[0026] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.
[0027] It should be noted that the terms "target," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.
[0028] Example 1
[0029] Figure 1 This is a flowchart of a text review method provided in Embodiment 1 of the present invention. This embodiment is applicable to the automated review of different fields in a target text to be reviewed using different review methods. The method can be executed by a text review device, which can be implemented in software and / or hardware, and is generally integrated into an electronic device. This electronic device can be a terminal device or a server device, as long as it can execute the text review function. The present invention does not limit the specific type of electronic device. Correspondingly, as... Figure 1 As shown, the method includes the following operations:
[0030] S110. Obtain the target text to be reviewed.
[0031] The target text to be reviewed can be any text that requires review. For example, the target text to be reviewed can include, but is not limited to, academic papers, policy documents, market analysis reports, and due diligence reports, as long as the text requires review. This embodiment of the invention does not limit the type of target text to be reviewed.
[0032] In this embodiment of the invention, texts requiring review can be used as target texts to be reviewed, thereby enabling automated review of the target texts.
[0033] S120. The target text to be reviewed is parsed to determine the target reference fields of the target text to be reviewed; wherein, the target reference fields include target public fields and target sensitive fields.
[0034] The target reference field can be a relevant field in the target text to be reviewed. For example, the target reference field can include, but is not limited to, publicly disclosed fields and sensitive fields; this embodiment of the invention does not limit the specific content of the target reference field. A publicly disclosed field can be a field in the target text to be reviewed that can be publicly disclosed. A sensitive field can be a field in the target text to be reviewed that cannot be publicly disclosed.
[0035] Accordingly, after obtaining the target text to be reviewed, regular expressions or target big data models can be used to parse the target text to identify the target public fields and target sensitive fields in the target text.
[0036] Understandably, the content of publicly disclosed and sensitive fields can differ depending on the type of the target document to be reviewed. In a specific example, a due diligence report is a document used in business transactions, investments, mergers and acquisitions, or other significant decision-making processes to comprehensively assess and review a target company or asset. In the context of increasingly stringent compliance requirements in the financial industry, supplier onboarding review is a crucial step in ensuring sound business operations and mitigating operational and compliance risks. A detailed and accurate due diligence report is the core basis for assessing supplier qualifications and identifying potential risks. Assuming the target document to be reviewed is a due diligence report, the publicly disclosed fields can include basic information such as the supplier's name, registered address, and business scope. Sensitive fields can be those in the due diligence report that involve trade secrets, personal privacy, or other sensitive information.
[0037] S130. The content of the target reference field is reviewed using a matching text review method according to the type of the target reference field.
[0038] Accordingly, after determining the target public fields and target sensitive fields of the target text to be reviewed, different text review methods can be used to automatically review the content of the target public fields and target sensitive fields of the target text to be reviewed.
[0039] For example, for publicly accessible target fields, a public internet channel can be used for review; for sensitive target fields, a non-public intranet channel can be used for review.
[0040] In summary, the text review method provided by this invention employs different text review approaches to review target public fields and target sensitive fields, thus more accurately meeting the review needs of different fields. This ensures both the compliance and accuracy of public information while effectively protecting the security and confidentiality of sensitive information. This differentiated review strategy not only improves text review efficiency but also reduces potential risks arising from improper information processing, thereby better meeting the needs of text review. Furthermore, the automated text review method avoids the problems of information verification delays and insufficient review coverage caused by relying on manual review.
[0041] This invention, through obtaining and parsing the target text to be reviewed, determines the target reference fields. These target reference fields include publicly accessible fields and sensitive fields. After determining the publicly accessible and sensitive fields, the content of the target reference fields is reviewed using a matching text review method based on their type. This solution addresses the problems of low review efficiency, delayed information verification, and insufficient review coverage in existing text review methods, improving both the efficiency and coverage of text review while ensuring the security of core data.
[0042] Example 2
[0043] Figure 2 This is a flowchart of a text review method provided in Embodiment 2 of the present invention. This embodiment is based on the above embodiment and is further specified. In this embodiment, several specific optional implementation methods are given for reviewing the content of the target reference field by applying a matching text review method according to the type of the target reference field. Correspondingly, such as Figure 2 As shown, the method in this embodiment may include:
[0044] S210. Obtain the target text to be reviewed.
[0045] S220. Parse the target text to be reviewed to determine the target reference fields of the target text to be reviewed; wherein, the target reference fields include target public fields and target sensitive fields.
[0046] Figure 3 This is a framework diagram of a text review system provided in Embodiment 2 of the present invention. In a specific example, such as Figure 3 As shown, a text review system may include a preprocessing module. The preprocessing module can use regular expressions and a target big data model to automatically parse the target text to be reviewed, dynamically identifying and marking target public fields and target sensitive fields in the target text.
[0047] S230. Determine whether the type of the target reference field is a target public field. If yes, execute S240a; otherwise, execute S240b.
[0048] S240a. Review the content of the target public field through an open Internet channel.
[0049] Among them, the Internet public channel can be a channel used to realize data transmission and communication in the Internet environment.
[0050] Specifically, after determining the target reference field of the text to be reviewed, the type of the target reference field can be determined. If the type of the target reference field is a target public field, the data transmission link of the Internet public channel can be activated, and the field content of the target public field can be reviewed through the Internet public channel.
[0051] In a specific example, such as Figure 3 As shown, the text review system may also include a routing decision module. This module can perform routing control based on field type labels generated by the preprocessing module: when a target public field is identified, the data transmission link of the public internet channel is activated; when a target sensitive field is identified, the data transmission link of the non-public intranet channel is activated. Simultaneously, the routing decision, through a hard-coded strategy, can achieve millisecond-level response, directing the target public field to the public internet channel for real-time retrieval and automated verification of external authoritative data sources, thereby ensuring the timeliness and accuracy of the target text to be reviewed. Furthermore, its logic circuitry and dual physical network cards adopt a direct connection design, ensuring path isolation.
[0052] In an optional embodiment of the present invention, the review of the field content of the target public field through an internet public channel may include: querying a multi-source database connected to the internet public channel based on the target public field to obtain query results for the target public field; calculating the text similarity between each target public field and the query results of the target public field; and generating a difference analysis result between each target public field and the query results of the target public field if the text similarity between each target public field and the query results of the target public field exceeds a preset similarity threshold.
[0053] The multi-source database can be multiple publicly accessible databases connected via a public internet channel. The query results for the target publicly accessible field can be results from the multi-source databases related to the content of the target publicly accessible field. Text similarity can be the similarity between the target publicly accessible field and the query results for the target publicly accessible field. The preset similarity threshold can be a pre-set threshold for text similarity. For example, the preset similarity threshold can be 0.7; this embodiment of the invention does not limit the specific value of the preset similarity threshold. The difference analysis results can be the results obtained by performing difference analysis between each target publicly accessible field and its associated query results.
[0054] In a specific example, such as Figure 3 As shown, the text review system may also include publicly available internet channels. These channels are used to verify the authenticity of publicly available information. They can connect to multiple external data sources, such as enterprise registration databases and credit reporting systems, and have a pre-configured list of trusted industry data sources. When reviewing the content of a target publicly available field through the internet channel, the primary model engine can first query the multi-source databases connected to the internet channel based on the target publicly available field to determine the content associated with that field in the databases—the query results for the target publicly available field.
[0055] After obtaining the query results for the target public fields, the text similarity between each target public field and its query results can be calculated. If the text similarity between each target public field and its query results exceeds a preset similarity threshold, it can be determined that the query result is related to the content of the target public field. Based on this, a difference analysis result can be generated based on the target public field and its query results, and field-level inconsistency annotations and corresponding data source supporting information can be output. This process operates in a physically isolated hardware environment, ensuring no data intrusion risk between external network connections and internal systems. The above solution, through a multi-source database connected via a publicly accessible internet channel, can quickly acquire a large amount of relevant data and perform real-time comparative analysis with the target public fields. This automated review method effectively reduces the time and workload required for manual review, thereby significantly improving review efficiency. Simultaneously, the multi-source database provides rich data sources, covering information from different dimensions and levels, which helps to more comprehensively and accurately identify differences and problems in the target text to be reviewed, thereby improving the accuracy of the review results. Furthermore, the difference analysis results of the target public fields enable reviewers to have a more comprehensive understanding of the actual situation of the target text to be reviewed, thereby improving the quality of decision-making.
[0056] S240b: Review the content of the target sensitive fields through a non-public channel on the internal network.
[0057] Among them, the non-public channel on the internal network can be a channel used to review the target text to be reviewed in the internal network environment.
[0058] Specifically, if the target reference field is a sensitive field, the data transmission link of the internal non-public channel can be activated, and the content of the sensitive field can be reviewed through the internal non-public channel. This solution, leveraging a dual-channel collaborative mechanism, not only achieves efficient review of the target text to be reviewed but also ensures the security of core data.
[0059] In an optional embodiment of the present invention, the review of the field content of the target sensitive field through a non-public channel on the intranet may include: performing comparative analysis based on the target sensitive field and the target text knowledge graph to generate a compliance issue list for the target sensitive field; calling the target large model to review the conflict statements of the field content of the target sensitive field to obtain the conflict review result of the target sensitive field; and reviewing the integrity of the field content of the target sensitive field based on the historical text database to obtain the integrity review result of the target sensitive field.
[0060] Specifically, the target text knowledge graph can be a knowledge graph associated with the compliance of the target text content. The compliance issue list can be a list used to record non-compliance in the target sensitive fields. The conflict review result can be the result obtained after reviewing the conflict statements of the target sensitive fields. The historical text repository can be a knowledge base used to store compliant texts of the same type as the target text to be reviewed. The integrity review result can be the result obtained after reviewing the integrity of the target sensitive fields.
[0061] In this embodiment of the invention, when reviewing the content of target sensitive fields through a non-public intranet channel, a target big data model can be used to analyze the target sensitive fields and the target text knowledge graph. This identifies fields within the target sensitive fields that do not meet compliance requirements, as well as compliance requirements of the knowledge graph not covered by the target sensitive fields, and generates a list of compliance issues for the target sensitive fields. Simultaneously, the target big data model can be used to review the content of the target sensitive fields for conflicting statements, generating conflict review results for the target sensitive fields. Furthermore, the completeness of the content of the target sensitive fields can be reviewed based on a historical text database to determine whether the target sensitive fields cover all required content, thus yielding a completeness review result for the target sensitive fields. This solution, by reviewing the content of target sensitive fields through a non-public intranet channel, avoids the risk of leakage that may occur when sensitive information is transmitted over the internet. Moreover, this automated review method can rapidly process massive amounts of sensitive field content, significantly improving the overall efficiency and coverage of the review.
[0062] In a specific example, assuming the target report to be reviewed is a due diligence report, the knowledge graph of the target text can be a policy knowledge graph, and the historical text library can be a historical report library storing historical compliance due diligence reports. The policy knowledge graph can be obtained by parsing the organization's internal rules and regulations, breaking down the clauses into quantifiable checkpoints, and establishing logical relationships between these nodes. For example... Figure 3 As shown, the text review system can also include a non-public internal network channel. This channel can include local knowledge bases such as a policy knowledge graph and a historical report database. Through the second major model engine, this channel can achieve three core functions: First, policy clause mapping: the non-public internal network channel can identify the correspondence between target sensitive fields and policy knowledge graph nodes, locate uncovered mandatory clauses, and fields that do not meet policy compliance requirements, thereby generating a list of policy compliance issues; second, logical contradiction detection: the non-public internal network channel can identify expression contradictions using the semantic analysis capabilities of the target large model; and third, structural completeness verification: the non-public internal network channel can refer to the structural coverage of historical reports of similar businesses and, in conjunction with internal policy requirements, assess the completeness of the due diligence report's content structure. Identifying target sensitive fields through the non-public internal network channel can significantly improve the completeness of the policy requirements coverage in the due diligence report and the accuracy of risk identification.
[0063] A target large-scale model can be a type of large-scale model, also known as a large-scale language model. This refers to a deep learning model trained on a large amount of relevant data (such as text data, speech data, or combined text and image data) capable of processing text sequences. It can generate natural language text or understand the meaning of language text. These models typically have billions of parameters. Large-scale language models can handle various natural language tasks, such as text classification, question answering, and dialogue, and have wide applications. The input to a large-scale language model is data, such as text data, speech data, or combined text and image data. The large-scale language model encodes the input data to obtain corresponding word vector representations, and then decodes the encoded word vectors, thus automatically processing the input data and obtaining the corresponding output data. For example, text can be input into a large-scale language model, which processes and predicts the input text, outputting the corresponding response text. By using a target large-scale model to automatically parse the target text to be reviewed, the target reference fields of the target text to be reviewed can be obtained, which can improve the review efficiency of the target text.
[0064] In an optional embodiment of the present invention, after reviewing the field content of the target reference field according to the type of the target reference field using a matching text review method, the method may further include: generating a review report of the target text to be reviewed based on the difference analysis results, the compliance issue list, the conflict review results, and the integrity review results; calculating the review score of the publicly accessible Internet channel based on the difference analysis results; calculating the review score of the privately accessible intranet channel based on the compliance issue list, the conflict review results, and the integrity review results; and calculating a comprehensive review score of the target text to be reviewed based on the review scores of the publicly accessible Internet channel and the privately accessible intranet channel.
[0065] Specifically, the review report for the target text to be reviewed can be a report obtained after a comprehensive review of the target text. The review score through the publicly accessible internet channel can be the result of scoring the discrepancy analysis results. The review score through the non-public intranet channel can be the result of comprehensively scoring the compliance issue list, conflict review results, and integrity review results. The comprehensive review score for the target text to be reviewed can be the result calculated from the review scores through the publicly accessible internet channel and the review scores through the non-public intranet channel.
[0066] In this embodiment of the invention, after reviewing the content of the target reference field using a matching text review method according to the type of the target reference field, the review results can be integrated and comprehensively analyzed to obtain a review report and a comprehensive review score for the target text to be reviewed. Specifically, the difference analysis results, compliance issue list, conflict review results, and integrity review results obtained above can be integrated to obtain a review report for the target text to be reviewed. Simultaneously, the review score for the publicly accessible internet channel can be calculated based on the difference analysis results, and the review score for the non-public intranet channel can be calculated based on the compliance issue list, conflict review results, and integrity review results. After obtaining the review scores for the publicly accessible internet channel and the non-public intranet channel, a weighted sum can be performed on the review scores for the publicly accessible internet channel and the non-public intranet channel, and the weighted sum result can be used as the comprehensive review score for the target text to be reviewed.
[0067] In a specific example, such as Figure 3 As shown, the text review system may also include a results fusion module. This module aggregates review results from both public internet channels and private intranet channels, generating a review report that includes external data verification conclusions and internal compliance ratings, thereby ensuring the comprehensiveness and integrity of the review process. Furthermore, this module can provide a comprehensive score for the review report, optimizing decision support and providing decision-makers with more accurate and comprehensive reference data.
[0068] In an optional embodiment of the present invention, after reviewing the field content of the target reference field using a matching text review method according to the type of the target reference field, the method may further include: obtaining real-time review feedback information during the review process of reviewing the field content of the target reference field using a matching text review method according to the type of the target reference field; and adjusting the text review method according to the real-time review feedback information.
[0069] Among them, real-time review feedback information can be feedback information during the process of reviewing the field content of the target reference field using different text review methods.
[0070] In this embodiment of the invention, after reviewing the content of the target reference field using a matching text review method based on the type of the target reference field, real-time review feedback information can be obtained during the text review process. This feedback information allows for adjustments to the text review method, thereby improving the accuracy of the text review. In a specific example, the real-time review feedback information may include, but is not limited to, whether the large model accurately classifies the target reference field and whether the large model can correctly determine the compliance of the target sensitive field based on the target text knowledge graph. This embodiment of the invention does not limit the specific content included in the real-time review feedback information.
[0071] Optionally, the public internet channel is located in a first memory region; the private intranet channel is located in a second memory region; wherein: the first memory region and the second memory region are isolated from each other at the network layer, and the public internet channel and the private intranet channel each occupy an independent processor core, and memory sharing access is prohibited between the processor core occupied by the public internet channel and the processor core occupied by the private intranet channel.
[0072] The first memory area can be used to store data related to publicly accessible internet channels. The second memory area can be used to store data related to non-public access within an intranet.
[0073] Specifically, the public internet channel and the private intranet channel employ a hardware resource isolation mechanism. The public internet channel operates in a first memory region, which only caches intermediate comparison results with external multi-source databases. The private intranet channel operates in a second memory region, which can be used to store the knowledge graph of the target text and parameters of the target large model. The first and second memory regions are physically partitioned through a memory controller. Simultaneously, the first and second memory regions are isolated from each other at the network layer, belonging to independent PCIe (Peripheral Component Interconnect Express) channels. Furthermore, the public internet channel and the private intranet channel each occupy independent processor cores, and memory sharing access is prohibited between the processor cores used by the public internet channel and the private intranet channel. This scheme strictly implements network isolation, memory isolation, and computational isolation between the public internet channel and the private intranet channel, ensuring that classified information is processed only within the secure intranet environment, thereby achieving zero leakage.
[0074] This invention, through obtaining and parsing the target text to be reviewed, determines the target reference fields. These target reference fields include publicly accessible fields and sensitive fields. After determining the target reference fields, their type can be assessed. If a target reference field is determined to be publicly accessible, its content is reviewed through a publicly accessible internet channel. If a target reference field is determined to be sensitive, its content is reviewed through a non-public intranet channel. This solution addresses the problems of low review efficiency, delayed information verification, and insufficient review coverage in existing text review methods, improving both the efficiency and coverage of text review while ensuring the security of core data.
[0075] In the technical solution disclosed herein, the information collected is information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of the relevant data all comply with the relevant laws, regulations and standards of the relevant countries and regions, necessary confidentiality measures have been taken, and they do not violate public order and good morals. Corresponding operation entry points are provided for users to choose to authorize or refuse.
[0076] It should be noted that, in this embodiment of the invention, a corresponding operation entry can be provided to the user, allowing the user to choose to agree to or reject the automated decision result; if the user chooses to reject, the process will proceed to the expert decision-making process.
[0077] It should be noted that any arrangement or combination of the technical features in the above embodiments also falls within the protection scope of this invention.
[0078] Example 3
[0079] Figure 4 This is a schematic diagram of a text review device provided in Embodiment 3 of the present invention, as shown below. Figure 4 As shown, the device includes: a target text acquisition module 310, a target reference field determination module 320, and a text review module 330, wherein:
[0080] The target text to be reviewed acquisition module 310 is used to acquire the target text to be reviewed.
[0081] The target reference field determination module 320 is used to parse the target text to be reviewed and determine the target reference fields of the target text to be reviewed; wherein, the target reference fields include target public fields and target sensitive fields.
[0082] The text review module 330 is used to review the field content of the target reference field using a matching text review method according to the type of the target reference field.
[0083] This invention, through obtaining and parsing the target text to be reviewed, determines the target reference fields. These target reference fields include publicly accessible fields and sensitive fields. After determining the publicly accessible and sensitive fields, the content of the target reference fields is reviewed using a matching text review method based on their type. This solution addresses the problems of low review efficiency, delayed information verification, and insufficient review coverage in existing text review methods, improving both the efficiency and coverage of text review while ensuring the security of core data.
[0084] Optionally, the text review module 330 is specifically used to: review the content of the target public field through a public Internet channel when the target reference field is determined to be the target public field; and review the content of the target sensitive field through a non-public intranet channel when the target reference field is determined to be the target sensitive field.
[0085] Optionally, the text review module 330 is further configured to: query the multi-source database connected to the Internet public channel according to the target public field to obtain the query results of the target public field; calculate the text similarity between each target public field and the query results of the target public field; and generate a difference analysis result between each target public field and the query results of the target public field if it is determined that the text similarity between each target public field and the query results of the target public field exceeds a preset similarity threshold.
[0086] Optionally, the text review module 330 is further configured to: perform comparative analysis based on the target sensitive field and the target text knowledge graph to generate a compliance issue list for the target sensitive field; call the target big model to review the conflict statements of the field content of the target sensitive field to obtain the conflict review result of the target sensitive field; and review the integrity of the field content of the target sensitive field based on the historical text library to obtain the integrity review result of the target sensitive field.
[0087] Optionally, the above apparatus may further include a review report generation module, used to generate a review report of the target text to be reviewed based on the difference analysis results, the compliance issue list, the conflict review results, and the integrity review results; calculate the review score of the publicly accessible internet channel based on the difference analysis results; calculate the review score of the non-public intranet channel based on the compliance issue list, the conflict review results, and the integrity review results; and calculate the comprehensive review score of the target text to be reviewed based on the review scores of the publicly accessible internet channel and the non-public intranet channel.
[0088] Optionally, the public internet channel is located in a first memory region; the private intranet channel is located in a second memory region; wherein: the first memory region and the second memory region are isolated from each other at the network layer, and the public internet channel and the private intranet channel each occupy an independent processor core, and memory sharing access is prohibited between the processor core occupied by the public internet channel and the processor core occupied by the private intranet channel.
[0089] The above-described text review device can execute the text review method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the method. Technical details not described in detail in this embodiment can be found in the text review method provided in any embodiment of the present invention.
[0090] Optionally, the above-mentioned device may further include a text review method adjustment module, used to: obtain real-time review feedback information during the process of reviewing the field content of the target reference field according to the type of the target reference field using a matching text review method; and adjust the text review method according to the real-time review feedback information.
[0091] Since the text review device described above is capable of executing the text review method in the embodiments of the present invention, those skilled in the art can understand the specific implementation and various variations of the text review device in this embodiment based on the text review method described in the embodiments of the present invention. Therefore, how the text review device implements the text review method in the embodiments of the present invention will not be described in detail here. Any device used by those skilled in the art to implement the text review method in the embodiments of the present invention falls within the scope of protection of this application.
[0092] Example 4
[0093] Figure 5 A schematic diagram of an electronic device 10 that can be used to implement embodiments of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.
[0094] like Figure 5As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 may also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.
[0095] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.
[0096] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as text review methods.
[0097] In some embodiments, the text review method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the text review method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the text review method by any other suitable means (e.g., by means of firmware).
[0098] Optionally, the text review method may include: obtaining the target text to be reviewed; parsing the target text to be reviewed to determine the target reference fields of the target text to be reviewed; wherein the target reference fields include target public fields and target sensitive fields; and reviewing the field content of the target reference fields using a matching text review method according to the type of the target reference fields.
[0099] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.
[0100] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.
[0101] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.
[0102] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).
[0103] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or computing systems that include middleware components (e.g., application servers), or computing systems that include frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.
[0104] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.
[0105] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.
[0106] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.
Claims
1. A text review method characterized by, include: Obtain the target text to be reviewed; The target text to be reviewed is parsed to determine the target reference fields of the target text to be reviewed; wherein, the target reference fields include target public fields and target sensitive fields; The content of the target reference field is reviewed using a matching text review method based on the type of the target reference field.
2. The method according to claim 1, characterized in that, The step of reviewing the content of the target reference field using a matching text review method based on the type of the target reference field includes: If the target reference field is determined to be the target public field, the field content of the target public field is reviewed through a public Internet channel; If the target reference field is determined to be the target sensitive field, the content of the target sensitive field is reviewed through a non-public channel on the internal network.
3. The method according to claim 2, characterized in that, The review of the field content of the target public field through a public internet channel includes: The query results for the target public field are obtained by querying the multi-source database connected to the Internet public channel based on the target public field. Calculate the text similarity between each of the target public fields and the query results of the target public fields; If the text similarity between each of the target public fields and the query results of the target public fields exceeds a preset similarity threshold, a difference analysis result between each of the target public fields and the query results of the target public fields is generated.
4. The method according to claim 2, characterized in that, The review of the target sensitive field content through a non-public intranet channel includes: A comparative analysis is performed on the target sensitive fields and the target text knowledge graph to generate a list of compliance issues for the target sensitive fields; The target large model is invoked to review the conflicting statements of the field content of the target sensitive field, and the conflict review result of the target sensitive field is obtained; The integrity of the target sensitive field content is reviewed based on the historical text database to obtain the integrity review result of the target sensitive field.
5. The method according to any one of claims 3 or 4, characterized in that, After reviewing the content of the target reference field using a matching text review method based on the type of the target reference field, the method further includes: A review report for the target text to be reviewed is generated based on the results of the difference analysis, the list of compliance issues, the results of the conflict review, and the results of the integrity review. The review score for the publicly accessible internet channel is calculated based on the results of the difference analysis. The review score for the non-public intranet channel is calculated based on the compliance issue list, the conflict review results, and the integrity review results. The comprehensive review score of the target text to be reviewed is calculated based on the review scores from the publicly accessible internet channel and the non-public intranet channel.
6. The method according to any one of claims 2-5, characterized in that: The publicly accessible internet channel is located in the first memory region; the private intranet channel is located in the second memory region; wherein: The first memory region and the second memory region are isolated from each other at the network layer, and the public Internet channel and the private intranet channel each occupy an independent processor core. Memory sharing access is prohibited between the processor core occupied by the public Internet channel and the processor core occupied by the private intranet channel.
7. The method according to claim 1, characterized in that, After reviewing the content of the target reference field using a matching text review method based on the type of the target reference field, the method further includes: Obtain real-time review feedback information during the process of reviewing the field content of the target reference field using a matching text review method based on the type of the target reference field; The text review method is adjusted based on the real-time review feedback information.
8. A text review device, characterized in that, include: The target text to be reviewed acquisition module is used to acquire the target text to be reviewed; The target reference field determination module is used to parse the target text to be reviewed and determine the target reference fields of the target text to be reviewed; wherein, the target reference fields include target public fields and target sensitive fields; The text review module is used to review the content of the target reference field using a matching text review method based on the type of the target reference field.
9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the text review method according to any one of claims 1-7.
10. A computer program product comprising a computer program / instructions, wherein, When the computer program / instructions are executed by the processor, they implement the text review method according to any one of claims 1-7.