Intelligent customization method and device for personalized service based on LLM security protection system

Abstract

The application discloses a kind of personalized service intelligent customization method and device based on LLM security protection system, by implementing the present application, after obtaining the customization demand information of target user, specific protection scene information and target demand details can be accurately determined, then by matching with historical protection scheme, the first protection scheme most suitable for target demand details is screened out, avoids the tedious customization process from zero, improves the determination efficiency of first protection scheme;Then, according to protection scene information and target demand details, scheme details adjustment and / or scheme verification can be carried out on the first protection scheme, through scheme details adjustment, fine customization can be carried out according to the special needs of users, improve the customization accuracy and practicability of second protection scheme;Meanwhile, the scheme verification can ensure that the second protection scheme after customization has no logical loopholes, is stable and reliable in operation, enhances the overall stability and reliability of security protection system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of security protection, and in particular to a personalized service intelligent customization method and device based on an LLM security protection system. BACKGROUND

[0002] With the rapid development of information technology, network security problems have become increasingly prominent, and various security protection systems have emerged to provide comprehensive and effective security protection for users. Among them, the large language model (LLM) security protection system, as an advanced network security solution, has received widespread attention and application due to its high flexibility and customizability.

[0003] However, in the use of existing LLM security protection systems, in combination with the diverse use needs of users, the system often needs to face the technical problems of high system use limitations and non-adaptation to different user needs. To solve this technical problem, the traditional solution is to manually configure various security parameters and strategies by users / service providers according to user needs. However, this approach not only consumes time and effort, but also easily leads to improper configuration due to the limitations of user expertise and experience, thereby affecting the security protection effect. Therefore, it is particularly important to provide a corresponding solution to the technical problems of high system application limitations and high system customization complexity for adapting to user needs when using the LLM security protection system in the prior art. SUMMARY

[0004] The present application provides a personalized service intelligent customization method and device based on an LLM security protection system, which can realize the intelligentization and customization of the security protection system, reduce the customization difficulty of the security protection system and improve the system customization efficiency, and broaden the use scenarios of the security protection system.

[0005] To solve the above technical problems, the present application discloses a personalized service intelligent customization method based on an LLM security protection system, which comprises:

[0006] detecting customization demand information of a target user for the LLM security protection system, and analyzing the customization demand information to obtain protection scenario information and target demand details for the LLM security protection system;

[0007] obtaining a plurality of historical protection schemes recorded in a database and matching the protection scenario information, each of the historical protection schemes comprising a plurality of protection sub-items, and each of the protection sub-items being used to adapt to one demand detail of a historical user;

[0008] According to the target requirement details, requirement details pairing is performed on all the historical protection schemes to obtain a first protection scheme with a highest target adaptation value with the target requirement details;

[0009] According to the protection scene information and the target requirement details, scheme customization is performed on the first protection scheme to obtain a second protection scheme corresponding to the first protection scheme, and a customization operation is performed on the LLM security protection system according to the second protection scheme; wherein the scheme customization comprises scheme details adjustment and / or scheme verification.

[0010] As an optional implementation, in the first aspect of the application, the protection scene information comprises a protection scene in which the customized requirement information is actually applied, and the protection scene comprises any one of a portal website, an e-commerce platform, a financial industry, and a game industry.

[0011] The target requirement details comprise a plurality of requirement customization matters matched with the protection scene, and the target requirement details further comprise a matter type corresponding to each of the requirement customization matters, and the matter type comprises a rule adjustment type or a configuration parameter adjustment type.

[0012] Furthermore, the plurality of historical protection schemes recorded in the database and matched with the protection scene information are obtained, comprising:

[0013] All the basic protection schemes recorded in the database are obtained, and according to a scene label of the protection scene in the protection scene information, first-level screening is performed on all the basic protection schemes to obtain a plurality of first-level screening protection schemes matched with the scene label.

[0014] According to a preset screening parameter, second-level screening is performed on all the first-level screening protection schemes to obtain a plurality of second-level screening protection schemes as the historical protection schemes; wherein the screening parameter comprises a scheme recording date and a scheme quantity; the scheme recording date is used to screen schemes within a target recording date from all the first-level screening protection schemes; and the scheme quantity is used to limit the quantity of all the second-level screening protection schemes.

[0015] As an optional implementation, in the first aspect of the application, according to the target requirement details, requirement details pairing is performed on all the historical protection schemes to obtain a first protection scheme with a highest target adaptation value with the target requirement details, comprising:

[0016] For each of the historical protection schemes, all the requirement customization matters corresponding to the target requirement details are matched with all the protection sub-items corresponding to the historical protection scheme to obtain first adaptation information of the target requirement details and the historical protection scheme, and the first adaptation information comprises a first adaptation value.

[0017] determining a user group classification of the target user and a corresponding user portrait according to the customized demand information; and extracting customized preference information adapted to the user group classification from the user portrait;

[0018] performing a scheme adaptation value evaluation on each of the historical protection schemes according to the customized preference information, to obtain second adaptation information corresponding to the target user and each of the historical protection schemes, the second adaptation information including a second adaptation value;

[0019] determining a first protection scheme with a highest target adaptation value corresponding to the target demand specification according to the first adaptation information and the second adaptation information corresponding to each of the historical protection schemes.

[0020] As an optional implementation, in the first aspect of the present application, the first adaptation information corresponding to each of the historical protection schemes further includes adaptation items and difference items between all the protection sub-items corresponding to the historical protection scheme and all the demand customization items corresponding to the target demand specification;

[0021] The determining of the first protection scheme with the highest target adaptation value corresponding to the target demand specification according to the first adaptation information and the second adaptation information corresponding to each of the historical protection schemes includes:

[0022] For each of the historical protection schemes, a target adaptation value corresponding to the customized demand information is calculated according to the first adaptation value corresponding to the historical protection scheme and a first preset weight coefficient corresponding thereto, and the second adaptation value and a second preset weight coefficient corresponding thereto;

[0023] adaptation specification information of the historical protection scheme is generated according to the adaptation items and the difference items corresponding to the historical protection scheme, and the adaptation specification information is added to the historical protection scheme;

[0024] The first protection scheme with the highest target adaptation value is selected from all the historical protection schemes.

[0025] As an optional implementation, in the first aspect of the present application, the performing of the scheme customization on the first protection scheme according to the protection scene information and the target demand specification to obtain a second protection scheme corresponding to the first protection scheme includes:

[0026] According to all the demand customization matters corresponding to the target requirement details, comparison is performed on all the protection sub-items included in the first protection scheme, to obtain corresponding similar matters, first difference matters and second difference matters, the first difference matters being matters remaining after all the demand customization matters are screened from the similar matters, and the second difference matters being matters remaining after all the protection sub-items are screened from the similar matters;

[0027] For the first protection scheme, the first difference matters are added, the similar matters are retained, and the second difference matters are deleted, to obtain a new first protection scheme, and it is determined that the first difference matters and the similar matters are all protection customization matters of the first protection scheme;

[0028] For each protection customization matter, matter filling is performed according to the matter type corresponding to the protection customization matter, to update the protection customization matter;

[0029] After it is determined that all the protection customization matters are updated, the new first protection scheme is determined as a second protection scheme.

[0030] As an optional implementation, in the first aspect of the present application, the matter filling performed for each protection customization matter according to the matter type corresponding to the protection customization matter comprises:

[0031] For each protection customization matter, when the matter type is the rule adjustment type, all target filling information matched with the protection customization matter is determined from a plurality of preset filling information, and the all target filling information matched with the protection customization matter is filled into the protection customization matter; wherein the plurality of preset filling information comprises access control rules, intrusion detection and protection rules, Web application firewall rules and other security rules;

[0032] When the matter type is the configuration parameter adjustment type, a state control instruction and a matter control parameter for the protection customization matter are generated, and the state control instruction and the matter control parameter of the protection customization matter are added to the protection customization matter;

[0033] Wherein, the state control instruction of each protection customization matter is used to control enabling or disabling the protection customization matter, and the matter control parameter of each protection customization matter is used to indicate the execution frequency of the control instruction corresponding to the matter control parameter.

[0034] As an optional implementation, in the first aspect of the present application, when the protection scenario is the e-commerce platform, all the protection sub-items corresponding to the e-commerce platform include one or more of payment security, user information protection, anti-fraud detection, and business failure monitoring and processing.

[0035] When the protection scenario is the e-commerce platform, all the protection sub-items corresponding to the e-commerce platform include one or more of payment security, user information protection, anti-fraud detection, and business failure monitoring and processing.

[0036] When the protection scenario is the financial industry, all the protection sub-items corresponding to the e-commerce platform include one or more of data encryption and transmission security, identity authentication and access control, and business continuity monitoring and failure handling.

[0037] When the protection scenario is the game industry, all the protection sub-items corresponding to the e-commerce platform include one or more of DDos attack protection, game data security, anti-cheating mechanism, and server load balancing.

[0038] The second aspect of the present application discloses an intelligent customization device for personalized services based on an LLM security protection system, which comprises:

[0039] A detection module for detecting customization demand information of a target user for an LLM security protection system;

[0040] A demand analysis module for analyzing the customization demand information to obtain protection scenario information and target demand details for the LLM security protection system;

[0041] An acquisition module for acquiring a plurality of historical protection schemes recorded in a database and matching the protection scenario information, each of the historical protection schemes comprising a plurality of protection sub-items, each of the protection sub-items being used to adapt to one demand detail of a historical user;

[0042] A pairing module for performing demand detail pairing on all the historical protection schemes according to the target demand details to obtain a first protection scheme with the highest target adaptation value corresponding to the target demand details;

[0043] A scheme customization module for performing scheme customization on the first protection scheme according to the protection scenario information and the target demand details to obtain a second protection scheme corresponding to the first protection scheme, and performing customization operation on the LLM security protection system according to the second protection scheme; wherein the scheme customization includes scheme detail adjustment and / or scheme verification.

[0044] As an optional implementation, in the second aspect of the present application, the protection scene information includes a protection scene to which the customized demand information is actually applied, and the protection scene includes any one of a portal website, an e-commerce platform, a financial industry, and a game industry.

[0045] The target demand details include a plurality of demand customization items matched with the protection scene, and the target demand details further include an item type corresponding to each of the demand customization items, and the item type includes a rule adjustment type or a configuration parameter adjustment type.

[0046] In addition, the acquisition module acquires a plurality of historical protection schemes recorded in the database and matched with the protection scene information in the following manner:

[0047] All basic protection schemes recorded in the database are acquired, and a first-level screening is performed on all the basic protection schemes according to a scene label of the protection scene in the protection scene information, to obtain a plurality of first-level screening protection schemes matched with the scene label.

[0048] A second-level screening is performed on all the first-level screening protection schemes according to a preset screening parameter, to obtain a plurality of second-level screening protection schemes as the historical protection schemes; wherein the screening parameter includes a scheme recording date and a scheme quantity; the scheme recording date is used to screen schemes within a target recording date from all the first-level screening protection schemes; and the scheme quantity is used to limit the quantity of all the second-level screening protection schemes.

[0049] As an optional implementation, in the second aspect of the present application, the pairing module performs demand details pairing on all the historical protection schemes according to the target demand details in the following manner to obtain a first protection scheme with a highest target adaptation value of the target demand details:

[0050] For each of the historical protection schemes, all demand customization items corresponding to the target demand details are matched with all protection sub-items corresponding to the historical protection scheme, to obtain first adaptation information of the target demand details and the historical protection scheme, and the first adaptation information includes a first adaptation value.

[0051] According to the customized demand information, a user group classification of the target user and a corresponding user portrait are determined, and customized preference information adapted to the user group classification is extracted from the user portrait.

[0052] According to the customized preference information, a scheme adaptation value evaluation is performed on each of the historical protection schemes, to obtain second adaptation information corresponding to each of the historical protection schemes and the target user, and the second adaptation information includes a second adaptation value.

[0053] determine, according to the first adaptation information and the second adaptation information corresponding to each of the historical protection schemes, a first protection scheme with a highest target adaptation value corresponding to the target requirement details.

[0054] As an optional implementation, in the second aspect of the present application, the first adaptation information corresponding to each of the historical protection schemes further includes adaptation items and difference items between all the protection sub-items corresponding to the historical protection scheme and all the requirement customization items corresponding to the target requirement details.

[0055] The pairing module determines, according to the first adaptation information and the second adaptation information corresponding to each of the historical protection schemes, a first protection scheme with a highest target adaptation value corresponding to the target requirement details.

[0056] For each of the historical protection schemes, a target adaptation value corresponding to the customization requirement information of the historical protection scheme is calculated according to the first adaptation value corresponding to the historical protection scheme and a first preset weight coefficient corresponding thereto, and the second adaptation value and a second preset weight coefficient corresponding thereto.

[0057] According to the adaptation items and the difference items corresponding to the historical protection scheme, adaptation detail information of the historical protection scheme is generated, and the adaptation detail information is added to the historical protection scheme.

[0058] A first protection scheme with a highest target adaptation value is selected from all the historical protection schemes.

[0059] As an optional implementation, in the second aspect of the present application, the scheme customization module performs scheme customization on the first protection scheme according to the protection scene information and the target requirement details, to obtain a second protection scheme corresponding to the first protection scheme.

[0060] According to all the requirement customization items corresponding to the target requirement details, comparison is performed on all the protection sub-items included in the first protection scheme, to obtain corresponding similar items, first difference items and second difference items. The first difference items are items remaining after all the requirement customization items are screened from the similar items. The second difference items are items remaining after all the protection sub-items are screened from the similar items.

[0061] For the first protection scheme, the first difference items are added, the similar items are retained, and the second difference items are deleted, to obtain a new first protection scheme. It is determined that the first difference items and the similar items are all protection customization items of the first protection scheme.

[0062] For each of the protection customization items, item filling is performed according to the item type corresponding to the protection customization item, to update the protection customization item;

[0063] After determining that all the protection customization items are updated, the new first protection scheme is determined as the second protection scheme.

[0064] As an optional implementation, in the second aspect of the present application, the manner in which the scheme customization module performs item filling for each of the protection customization items according to the item type corresponding to the protection customization item specifically comprises:

[0065] For each of the protection customization items, when the item type is the rule adjustment type, all target filling information matched with the protection customization item is determined from a plurality of preset filling information, and all target filling information matched with the protection customization item is filled into the protection customization item; wherein the plurality of preset filling information comprises access control rules, intrusion detection and protection rules, Web application firewall rules and other security rules;

[0066] When the item type is the configuration parameter adjustment type, a state control instruction and an item control parameter for the protection customization item are generated; and the state control instruction and the item control parameter of the protection customization item are added to the protection customization item;

[0067] Wherein, the state control instruction of each of the protection customization items is used to control enabling or disabling the protection customization item; and the item control parameter of each of the protection customization items is used to indicate the execution frequency of the control instruction corresponding to the item control parameter.

[0068] As an optional implementation, in the second aspect of the present application, when the protection scenario is the portal website, all the protection sub-items corresponding to the portal website comprise one or more of DDos attack protection, SQL injection protection, cross-site scripting attack protection, data backup and recovery;

[0069] When the protection scenario is the e-commerce platform, all the protection sub-items corresponding to the portal website comprise one or more of payment security, user information protection, anti-fraud detection, business fault monitoring and processing;

[0070] When the protection scenario is the financial industry, all the protection sub-items corresponding to the portal website comprise one or more of data encryption and transmission security, identity authentication and access control, business continuity monitoring and fault handling;

[0071] When the protection scenario is the game industry, all the protection sub-items corresponding to the portal website include one or more of DDos attack protection, game data security, anti-cheating mechanism, and server load balancing.

[0072] The third aspect of the present application discloses another kind of personalized service intelligent customization device based on LLM security protection system, the device comprises:

[0073] Memory storing executable program code;

[0074] Processor coupled with the memory;

[0075] The processor calls the executable program code stored in the memory to execute the personalized service intelligent customization method based on the LLM security protection system disclosed in the first aspect of the present application.

[0076] The fourth aspect of the present application discloses a computer storage medium, the computer storage medium stores computer instructions, when the computer instructions are called, the personalized service intelligent customization method based on the LLM security protection system disclosed in the first aspect of the present application is executed.

[0077] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0078] In the embodiment of the present application, a personalized service intelligent customization method based on an LLM security protection system is provided, which comprises the following steps: detecting customization demand information of a target user for the LLM security protection system, and analyzing the customization demand information to obtain protection scene information and target demand details for the LLM security protection system; obtaining a plurality of historical protection schemes recorded in a database and matched with the protection scene information, each historical protection scheme comprising a plurality of protection sub-items, each protection sub-item being used to adapt to one demand detail of a historical user; performing demand detail matching on all historical protection schemes according to the target demand details to obtain a first protection scheme with the highest target adaptation value for the target demand details; performing scheme customization on the first protection scheme according to the protection scene information and the target demand details to obtain a second protection scheme corresponding to the first protection scheme; and performing customization operation on the LLM security protection system according to the second protection scheme; wherein the scheme customization comprises scheme detail adjustment and / or scheme verification. It can be seen that, after obtaining the customization demand information of the target user, the specific protection scene information and the target demand details can be accurately determined, and the first protection scheme most close to the target demand details can be quickly screened out by matching with the historical protection schemes, thereby avoiding the cumbersome customization process from scratch and being conducive to improving the determination efficiency of the first protection scheme; further, after the first protection scheme is determined, the first protection scheme can be subjected to scheme detail adjustment and / or scheme verification according to the protection scene information and the target demand details, and through the scheme detail adjustment, the first protection scheme can be finely customized according to the special demand of the user, thereby being conducive to improving the customization accuracy and practicability of the second protection scheme; meanwhile, the scheme verification can ensure that the second protection scheme after customization has no logical loopholes and is stable and reliable in operation, thereby enhancing the overall stability and reliability of the security protection system. BRIEF DESCRIPTION OF DRAWINGS

[0079] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort.

[0080] Figure 1 is a flowchart of a personalized service intelligent customization method based on an LLM security protection system disclosed by the embodiment of the present application;

[0081] Figure 2 is a flowchart of another personalized service intelligent customization method based on an LLM security protection system disclosed by the embodiment of the present application;

[0082] Figure 3is a structure schematic view of a personalized service intelligent customization device based on an LLM security protection system according to an embodiment of the present application.

[0083] Figure 4 is a structure schematic view of another personalized service intelligent customization device based on an LLM security protection system according to an embodiment of the present application. DETAILED DESCRIPTION

[0084] In order to make the personnel in the art better understand the present application scheme, the technical scheme in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0085] The terms "first", "second", and the like in the specification of the present application and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, device, product, or end including a series of steps or units is not limited to the listed steps or units, but can optionally include steps or units not listed, or can optionally include other steps or units inherent to the process, method, product, or end.

[0086] In this document, the reference to "embodiments" means that the specific features, structures, or characteristics described in connection with the embodiments can be included in at least one embodiment of the present application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it mutually exclusive or alternative to other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0087] The application discloses a personalized service intelligent customization method and device based on an LLM security protection system.

[0088] Embodiment one

[0089] Please refer to Figure 1 , Figure 1 is a flowchart of a personalized service intelligent customization method based on an LLM security protection system. Among them, Figure 1 The personalized service intelligent customization method based on the LLM security protection system described can be applied to a personalized service intelligent customization device based on the LLM security protection system, and the application embodiments are not limited. As shown in Figure 1 The personalized service intelligent customization method based on the LLM security protection system can include the following operations:

[0090] 101, detecting customization demand information of a target user for the LLM security protection system, and analyzing the customization demand information to obtain protection scene information and target demand details for the LLM security protection system.

[0091] 102, obtaining a plurality of historical protection schemes recorded in a database and matched with the protection scene information, each historical protection scheme including a plurality of protection sub-items, and each protection sub-item being used to adapt to a demand detail of a historical user.

[0092] 103, performing demand detail pairing on all historical protection schemes according to the target demand details to obtain a first protection scheme with the highest target adaptation value of the target demand details.

[0093] 104, performing scheme customization on the first protection scheme according to the protection scene information and the target demand details to obtain a second protection scheme corresponding to the first protection scheme, and performing customization operation on the LLM security protection system according to the second protection scheme.

[0094] In the embodiment of the present application, the scheme customization includes scheme detail adjustment and / or scheme verification.

[0095] It can be seen that the implementation Figure 1 The described personalized service intelligent customization method based on the LLM security protection system can accurately determine the specific protection scene information and target demand details after obtaining the customization demand information of the target user, and then quickly filter out the first protection scheme closest to the target demand details by matching with the historical protection schemes, thereby avoiding the cumbersome customization process from scratch and improving the determination efficiency of the first protection scheme. Furthermore, after determining the first protection scheme, the scheme detail adjustment and / or scheme verification can be performed on the first protection scheme according to the protection scene information and the target demand details. Through the scheme detail adjustment, the fine customization can be performed according to the special needs of the user, thereby improving the customization accuracy and practicality of the second protection scheme. At the same time, the scheme verification can ensure that the second protection scheme after customization has no logical loopholes and is stable and reliable in operation, thereby enhancing the overall stability and reliability of the security protection system.

[0096] In an optional embodiment, the protection scene information includes a protection scene in which the customization demand information is actually applied, and the protection scene includes any one of a portal website, an e-commerce platform, a financial industry, and a game industry.

[0097] The target demand details include a plurality of demand customization items matched with the protection scene, and the target demand details further include an item type corresponding to each demand customization item, and the item type includes a rule adjustment type or a configuration parameter adjustment type.

[0098] In addition, the manner of obtaining the plurality of historical protection schemes recorded in the database and matched with the protection scene information in step 102 specifically includes:

[0099] All basic protection schemes recorded in the database are obtained, and according to the scene label of the protection scene in the protection scene information, a first-level screening is performed on all basic protection schemes to obtain a plurality of first-level screening protection schemes matched with the scene label.

[0100] According to a preset screening parameter, a second-level screening is performed on all first-level screening protection schemes to obtain a plurality of second-level screening protection schemes as historical protection schemes. The screening parameter includes a scheme record date and a scheme quantity. The scheme record date is used to screen the schemes within the target record date from all first-level screening protection schemes. The scheme quantity is used to limit the number of all second-level screening protection schemes.

[0101] In the embodiment of the present application, by explicitly defining the target user's actual application protection scene (such as a portal website, an e-commerce platform, a financial industry, a game industry, etc.) of the LLM security protection system, this detailed protection scene division helps to more accurately understand the user's actual needs, so as to select or customize a protection scheme that is more suitable for the application scene; at the same time, by explicitly defining the matter type of each demand customization matter, the related parameters or rules in the protection scheme can be more targetedly adjusted or configured, thereby improving the accuracy and effectiveness of the customized scheme.

[0102] It can be seen that in this optional embodiment, when obtaining the historical protection scheme, a multi-level screening strategy is adopted, and by taking the scene label of the protection scene information as the first level screening condition, the screening range can be quickly narrowed, and the basic protection scheme matching the scene label can be accurately obtained from the database, ensuring the pertinence and relevance of the subsequent customization process, and also helping to improve the screening efficiency; then, all first-level screening protection schemes are subjected to secondary screening according to the preset screening parameters (such as scheme record date, scheme quantity, etc.), by limiting the scheme record date, it can be ensured that the selected scheme has certain timeliness and applicability; and by limiting the number of schemes, information overload can be avoided while ensuring diversity, improving the accuracy and efficiency of the subsequent customization process.

[0103] In another optional embodiment, the manner of step 104 of performing scheme customization on the first protection scheme according to the protection scene information and the target demand details to obtain a second protection scheme corresponding to the first protection scheme specifically includes:

[0104] According to all demand customization matters corresponding to the target demand details, the comparison is performed on all protection sub-items included in the first protection scheme, to obtain corresponding similar matters, first difference matters and second difference matters, the first difference matters being the matters remaining after the similar matters are excluded from all demand customization matters, and the second difference matters being the matters remaining after the similar matters are excluded from all protection sub-items;

[0105] For the first protection scheme, the first difference matters are added, the similar matters are retained, and the second difference matters are deleted, to obtain a new first protection scheme, and it is determined that the first difference matters and the similar matters are all protection customization matters of the first protection scheme;

[0106] For each protection customization matter, matter filling is performed according to the matter type corresponding to the protection customization matter, to update the protection customization matter;

[0107] After it is determined that all protection customization matters are updated, the new first protection scheme is determined as the second protection scheme.

[0108] It can be seen that in the optional embodiment, by adding the first difference item and retaining the same type item, while deleting the second difference item, the scheme realizes accurate adjustment of the first protection scheme to make it more consistent with the specific requirements in the target requirement details. Further, for the new first protection scheme, an item filling process is set for each protection customization item, which improves the generation accuracy of the final second protection scheme.

[0109] In yet another optional embodiment, the manner of filling the item for each protection customization item according to the item type corresponding to the protection customization item specifically includes:

[0110] For each protection customization item, when the item type is a rule adjustment type, all target filling information matching the protection customization item is determined from a plurality of preset filling information, and all target filling information matching the protection customization item is filled into the protection customization item; wherein the plurality of preset filling information includes access control rules, intrusion detection and protection rules, Web application firewall rules and other security rules;

[0111] When the item type is a configuration parameter adjustment type, a state control instruction and an item control parameter for the protection customization item are generated; and the state control instruction and the item control parameter of the protection customization item are added to the protection customization item;

[0112] Wherein, the state control instruction of each protection customization item is used to control enabling or disabling the protection customization item; and the item control parameter of each protection customization item is used to indicate the execution frequency of the control instruction corresponding to the item control parameter.

[0113] In the optional embodiment, the other security rules include sensitive data protection, log recording and auditing, regular backup and recovery, and security policy update. Wherein, the sensitive data protection is used to classify and label sensitive data, implement data encryption technology, and ensure the confidentiality and integrity of data. Log recording and auditing are used to record system security events and operation logs, facilitating subsequent security auditing and troubleshooting. Regular backup and recovery are used to regularly backup important data and test the recovery ability of backup data to ensure the reliability and availability of data. Security policy update is used to regularly update the security policy of the system according to the latest security threat and vulnerability information, to improve the security protection capability of the system.

[0114] In the optional embodiment, it should be noted that when the protection scenario is a portal website, all protection sub-items corresponding to the portal website include one or more of DDos attack protection, SQL injection protection, cross-site scripting attack protection, and data backup and recovery.

[0115] When the protection scenario is an e-commerce platform, all protection sub-items corresponding to the portal website include one or more of payment security, user information protection, fraud detection, business failure monitoring and handling;

[0116] When the protection scenario is the financial industry, all protection sub-items corresponding to the portal website include one or more of data encryption and transmission security, identity authentication and access control, business continuity monitoring and failure handling;

[0117] When the protection scenario is the game industry, all protection sub-items corresponding to the portal website include one or more of DDos attack protection, game data security, anti-cheat mechanism, and server load balancing.

[0118] It can be seen that in this optional embodiment, an item filling mechanism for protection customization items based on different item types is set, wherein for the rule adjustment type, the target filling information matched with the specific protection customization item is determined from the plurality of preset filling information and filled into the item, that is, the corresponding security rules can be flexibly adjusted and applied according to different security requirements and environments, and this mechanism ensures that the security protection measures can accurately target specific threats or vulnerabilities, improving the pertinence and effectiveness of security protection; for the configuration parameter adjustment type, the state control instruction and the item control parameter for the specific protection customization item are generated and added to the item, the system allows the user to enable or disable certain protection customization items as needed and set the corresponding execution frequency, so that the user can flexibly and dynamically adjust the security protection strategy according to the actual security situation and resource allocation, to optimize performance and resource utilization, that is, it is beneficial to improve the use flexibility and reliability of the subsequent generated customized second protection scheme.

[0119] In another optional embodiment, the manner of step 104 of performing scheme customization on the first protection scheme according to the protection scenario information and the target requirement details to obtain the second protection scheme corresponding to the first protection scheme specifically further includes:

[0120] Before determining the new first protection scheme as the second protection scheme, performing trial customization on the LLM security protection system according to all protection customization items included in the new first protection scheme to obtain a test protection system corresponding to the LLM security protection system;

[0121] Performing item running test on the test protection system according to each requirement customization item corresponding to the target requirement details to obtain running test information corresponding to each requirement customization item;

[0122] According to the customization requirement information, determining the customization effect of the target user on the LLM security protection system;

[0123] For each piece of running test information, the running test information is compared with pre-set standard test information for the running test information to obtain a first test result of the running test information;

[0124] Determining whether all the first test results indicate that there is an unqualified test result; when it is determined that there is an unqualified test result among all the first test results, performing a first item adjustment on the demand customization item corresponding to the unqualified test result according to the standard test information corresponding to the unqualified test result;

[0125] When it is determined that there is no unqualified test result among all the first test results, performing effect comparison on all the running test information according to the customized effect to obtain a second test result for each piece of running test information;

[0126] Determine whether there is any substandard test result that does not match the customized effect among all the second test results, and when it is determined that there is no substandard test result that does not match the customized effect among all the second test results, trigger the execution of the above-mentioned determination of the new first protection scheme as the second protection scheme;

[0127] When it is determined that there is a substandard test result among all the second test results that does not match the customized effect, the customized effect is used as a benchmark, and the second item adjustment is performed on the customized item corresponding to the substandard test result until the effect similarity between the adjusted substandard test result and the customized effect is higher than the standard similarity value, triggering the execution of the above-mentioned determination of the new first protection plan as the second protection plan.

[0128] It can be seen that in this optional embodiment, a verification mechanism is set up for the new first protection plan. By performing trial customization and trial operation on each protection customization item in the new first protection plan, and performing a step-by-step comparison of the test results with the standard value (standard test information) and the user's ideal effect (customization effect), after determining that each required customization item matches the standard value and the user's ideal effect, the new first protection plan is determined to be the second protection plan; the thoroughness, accuracy and reliability of the determination of the second protection plan are improved.

[0129] Example 2

[0130] See also Figure 2 , Figure 2 This is a flow chart of another method for intelligent customization of personalized services based on the LLM security protection system disclosed in an embodiment of the present invention. Figure 2 The described personalized service intelligent customization method based on the LLM security protection system can be applied to the personalized service intelligent customization device based on the LLM security protection system, and the embodiment of the present invention does not limit it. Figure 2As shown, the personalized service intelligent customization method based on the LLM security protection system can include the following operations:

[0131] 201, detecting the customization demand information of the target user for the LLM security protection system, and analyzing the customization demand information to obtain protection scene information and target demand details for the LLM security protection system.

[0132] 202, obtaining a plurality of historical protection schemes recorded in the database and matching the protection scene information, each historical protection scheme including a plurality of protection sub-items, and each protection sub-item being used to adapt to one demand detail of a historical user.

[0133] 203, for each historical protection scheme, matching all demand customization items corresponding to the target demand details with all protection sub-items corresponding to the historical protection scheme to obtain first adaptation information of the target demand details and the historical protection scheme.

[0134] In the embodiment of the application, the first adaptation information includes a first adaptation value.

[0135] In the embodiment of the application, by calculating the first adaptation value, the matching degree of each historical protection scheme and the target demand details can be quantitatively evaluated, providing a reliable basis for subsequent screening and customization of protection schemes.

[0136] 204, according to the customization demand information, determining the user group classification of the target user and the corresponding user portrait, and extracting customization preference information adapting to the user group classification from the user portrait.

[0137] In the embodiment of the application, by determining the user group classification of the target user and the corresponding user portrait, the content of the customization scheme is further refined in the manner of extracting customization preference information adapting to the user group classification from the user portrait, so that it is more in line with the individual needs of the target user.

[0138] 205, according to the customization preference information, performing scheme adaptation value evaluation on each historical protection scheme to obtain second adaptation information corresponding to the target user and each historical protection scheme.

[0139] In the embodiment of the application, the second adaptation information includes a second adaptation value.

[0140] In the embodiment of the application, by calculating the second adaptation value, the matching degree of the target user and each historical protection scheme can be quantitatively reflected, so as to ensure that the finally selected first protection scheme can meet the individual needs of the target user to the greatest extent.

[0141] 206, according to the first adaptation information and the second adaptation information corresponding to each historical protection scheme, determining a first protection scheme with the highest target adaptation value of the target demand details.

[0142] In the embodiment of the present application, the final first protection scheme is determined by considering the first adaptation information and the second adaptation information. This comprehensive evaluation method avoids the one-sidedness and inaccuracy that may be caused by a single matching method, and is beneficial to improve the screening accuracy of the first protection scheme.

[0143] 207. According to the protection scene information and the target requirement details, a scheme customization is performed on the first protection scheme to obtain a second protection scheme corresponding to the first protection scheme; and a customization operation is performed on the LLM security protection system according to the second protection scheme.

[0144] In the embodiment of the present application, other descriptions of steps 201-202 and step 207 can refer to other specific descriptions of steps 101-102 and step 104 in Embodiment One, and the present embodiment will not be described again.

[0145] It can be seen that, in the implementation Figure 2 The described personalized service intelligent customization method based on the LLM security protection system calculates the first adaptation value and the second adaptation value respectively, in a quantitative manner, to accurately evaluate the matching degree of each historical protection scheme and the target requirement details, and the matching degree of the target user and each historical protection scheme, thereby improving the scientific nature of the subsequent determination of the first protection scheme. In addition, the determination of the first protection scheme is performed by introducing the user portrait and comprehensively considering the multi-dimensional scheme adaptation value, which is beneficial to improve the determination accuracy and reliability of the first protection scheme.

[0146] In an optional embodiment, the first adaptation information corresponding to each historical protection scheme further includes adaptation items and difference items between all protection sub-items corresponding to the historical protection scheme and all requirement customization items corresponding to the target requirement details;

[0147] The above method of determining the first protection scheme with the highest target adaptation value according to the first adaptation information and the second adaptation information corresponding to each historical protection scheme includes:

[0148] For each historical protection scheme, the target adaptation value of the historical protection scheme and the customization requirement information is calculated according to the first adaptation value corresponding to the historical protection scheme and the first preset weight coefficient corresponding thereto, and the second adaptation value and the second preset weight coefficient corresponding thereto.

[0149] According to the adaptation items and the difference items corresponding to the historical protection scheme, adaptation detail information of the historical protection scheme is generated; and the adaptation detail information is added to the historical protection scheme;

[0150] The first protection scheme with the highest target adaptation value is selected from all historical protection schemes.

[0151] In this optional embodiment, by listing the adaptation items and the difference items, it can be clearly understood in which aspects the historical protection scheme meets the target requirement details and in which aspects there are differences with the target requirement details. The detailed adaptation information helps to accurately adjust the subsequent customization process, ensuring that the finally generated protection scheme can meet the target requirements to the greatest extent.

[0152] It can be seen that in this optional embodiment, by comprehensively considering the adaptation information of multiple dimensions, the historical protection scheme that best matches the target requirement details can be quickly and accurately screened out, which is beneficial to reduce unnecessary iteration and adjustment process and improve customization efficiency. Further, by displaying the adaptation detail information, the transparency and credibility of the customization process can be enhanced, and the viewing convenience and transparency of the first protection scheme for the target user / scheme customizer can be improved.

[0153] Embodiment three

[0154] Please refer to Figure 3 , Figure 3 is a structure diagram of an intelligent customization device for personalized services based on an LLM security protection system according to an embodiment of the present application. The intelligent customization device for personalized services based on an LLM security protection system can be an intelligent customization terminal, device, system or server based on an LLM security protection system. The server can be a local server, a remote server or a cloud server (also known as a cloud server). When the server is not a cloud server, the non-cloud server can be connected to the cloud server for communication. The present application does not limit the embodiment. As shown in Figure 3 The intelligent customization device for personalized services based on an LLM security protection system can include a detection module 301, a requirement analysis module 302, an acquisition module 303, a matching module 304 and a scheme customization module 305, wherein:

[0155] The detection module 301 is configured to detect the customization requirement information of the target user for the LLM security protection system.

[0156] The requirement analysis module 302 is configured to analyze the customization requirement information to obtain protection scenario information and target requirement details for the LLM security protection system.

[0157] The acquisition module 303 is configured to acquire a plurality of historical protection schemes recorded in the database and matched with the protection scenario information. Each historical protection scheme includes a plurality of protection sub-items, and each protection sub-item is used to adapt to one requirement detail of a historical user.

[0158] The pairing module 304 is configured to perform demand specification pairing on all historical protection schemes according to the target demand specification, to obtain a first protection scheme with the highest target adaptation value with the target demand specification.

[0159] The scheme customization module 305 is configured to perform scheme customization on the first protection scheme according to the protection scene information and the target demand specification, to obtain a second protection scheme corresponding to the first protection scheme, and perform customization operation on the LLM security protection system according to the second protection scheme; wherein the scheme customization includes scheme specification adjustment and / or scheme verification.

[0160] It can be seen that the implementation of the described LLM security protection system-based personalized service intelligent customization apparatus can accurately determine the specific protection scene information and the target demand specification after obtaining the customization demand information of the target user, and quickly screen out the first protection scheme closest to the target demand specification by matching with the historical protection schemes, thereby avoiding the cumbersome customization process from scratch and being conducive to improving the determination efficiency of the first protection scheme. Figure 3 Further, after determining the first protection scheme, the first protection scheme can be adjusted in scheme specification and / or verified in scheme according to the protection scene information and the target demand specification. Through the adjustment in scheme specification, the first protection scheme can be finely customized according to the special demand of the user, which is conducive to improving the customization accuracy and practicability of the second protection scheme. Meanwhile, the verification in scheme can ensure that the second protection scheme after customization has no logical loopholes and is stable and reliable in operation, thereby enhancing the overall stability and reliability of the security protection system.

[0161] In an optional embodiment, the protection scene information includes a protection scene in which the customization demand information is actually applied, and the protection scene includes any one of a portal website, an e-commerce platform, a financial industry, and a game industry.

[0162] The target demand specification includes a plurality of demand customization items matched with the protection scene, and the target demand specification further includes an item type corresponding to each demand customization item, and the item type includes a rule adjustment type or a configuration parameter adjustment type.

[0163] The acquisition module 303 acquires a plurality of historical protection schemes recorded in the database and matched with the protection scene information in the following manner:

[0164] All basic protection schemes recorded in the database are acquired, and first-level screening is performed on all the basic protection schemes according to the scene label of the protection scene in the protection scene information, to obtain a plurality of first-level screening protection schemes matched with the scene label.

[0165] According to the preset screening parameter, the secondary screening is performed on all the first-level screening protection schemes, and a plurality of secondary screening protection schemes are obtained as historical protection schemes; wherein, the screening parameter includes scheme record date and scheme quantity; the scheme record date is used for screening the schemes within the target record date from all the first-level screening protection schemes; and the scheme quantity is used for limiting the quantity of all the secondary screening protection schemes.

[0166] It can be seen that in the optional embodiment, when the historical protection scheme is acquired, a multi-level screening strategy is adopted, the scene tag of the protection scene information is taken as the first-level screening condition, the screening range can be quickly narrowed, the basic protection scheme matched with the scene tag is accurately acquired from the database, the pertinence and relevance of the subsequent customization process are ensured, and the screening efficiency is improved; then, according to the preset screening parameter (such as the scheme record date, the scheme quantity and the like), the secondary screening is performed on all the first-level screening protection schemes, by limiting the scheme record date, it can be ensured that the selected scheme has certain timeliness and applicability; and by limiting the scheme quantity, the information overload can be avoided while the diversity is ensured, and the accuracy and efficiency of the subsequent customization process are improved.

[0167] In another optional embodiment, the pairing module 304 performs demand detail pairing on all the historical protection schemes according to the target demand details, and the manner of obtaining the first protection scheme with the highest target adaptation value with the target demand details specifically includes:

[0168] For each historical protection scheme, all the demand customization items corresponding to the target demand details are matched with all the protection sub-items corresponding to the historical protection scheme, first adaptation information of the target demand details and the historical protection scheme is obtained, and the first adaptation information includes a first adaptation value;

[0169] According to the customization demand information, the user group classification of the target user and the corresponding user portrait are determined; and the customization preference information of the adaptation user group classification is extracted from the user portrait;

[0170] According to the customization preference information, the scheme adaptation value evaluation is performed on each historical protection scheme, and second adaptation information corresponding to the target user and each historical protection scheme is obtained, the second adaptation information includes a second adaptation value;

[0171] According to the first adaptation information and the second adaptation information corresponding to each historical protection scheme, the first protection scheme with the highest target adaptation value with the target demand details is determined.

[0172] It can be seen that in the optional embodiment, by calculating the first adaptation value and the second adaptation value respectively, the matching degree of each historical protection scheme and the target demand details and the matching degree of the target user and each historical protection scheme are accurately evaluated in a quantitative manner, and the scientificity of subsequent determination of the first protection scheme is improved. In addition, by introducing the user portrait and comprehensively considering the multi-dimensional scheme adaptation value, the determination of the first protection scheme is beneficial to improve the accuracy and reliability of the determination of the first protection scheme.

[0173] In yet another optional embodiment, the first adaptation information corresponding to each historical protection scheme further includes adaptation items and difference items between all protection sub-items corresponding to the historical protection scheme and all demand customization items corresponding to the target demand details.

[0174] The pairing module 304 determines the first protection scheme with the highest target adaptation value to the target demand details according to the first adaptation information and the second adaptation information corresponding to each historical protection scheme. The specific manner includes:

[0175] For each historical protection scheme, the target adaptation value of the historical protection scheme to the customization demand information is calculated according to the first adaptation value corresponding to the historical protection scheme and the first preset weight coefficient corresponding thereto, and the second adaptation value and the second preset weight coefficient corresponding thereto.

[0176] According to the adaptation items and the difference items corresponding to the historical protection scheme, the adaptation details information of the historical protection scheme is generated, and the adaptation details information is added to the historical protection scheme.

[0177] The first protection scheme with the highest target adaptation value is selected from all historical protection schemes.

[0178] It can be seen that in the optional embodiment, by comprehensively considering the adaptation information of multiple dimensions, the historical protection scheme most matched with the target demand details can be quickly and accurately screened out, which is beneficial to reduce unnecessary iteration and adjustment process and improve customization efficiency. Further, by displaying the adaptation details information, the transparency and credibility of the customization process can be enhanced, and the viewing convenience and transparency of the target user / scheme customizer to the first protection scheme can be improved.

[0179] In another optional embodiment, the scheme customization module 305 performs scheme customization on the first protection scheme according to the protection scene information and the target demand details to obtain the second protection scheme corresponding to the first protection scheme. The specific manner includes:

[0180] According to all the demand customization matters corresponding to the target requirement details, comparison is performed on all the protection sub-items included in the first protection scheme, to obtain corresponding similar matters, first difference matters and second difference matters, the first difference matters being matters remaining after the similar matters are filtered out from all the demand customization matters, and the second difference matters being matters remaining after the similar matters are filtered out from all the protection sub-items;

[0181] For the first protection scheme, the first difference matters are added, the similar matters are retained, and the second difference matters are deleted, to obtain a new first protection scheme, and it is determined that the first difference matters and the similar matters are all protection customization matters of the first protection scheme;

[0182] For each protection customization matter, matter filling is performed according to the matter type corresponding to the protection customization matter, to update the protection customization matter;

[0183] After it is determined that all the protection customization matters are updated, the new first protection scheme is determined as the second protection scheme.

[0184] It can be seen that in the optional embodiment, by adding the first difference matters and retaining the similar matters, while deleting the second difference matters, the scheme realizes accurate adjustment of the first protection scheme, so that it is more in line with the specific requirements in the target requirement details; further, for the new first protection scheme, matter filling process is set for each protection customization matter, to improve the generation accuracy of the final obtained second protection scheme.

[0185] In yet another optional embodiment, the scheme customization module 305 performs matter filling for each protection customization matter according to the matter type corresponding to the protection customization matter in the following manner:

[0186] For each protection customization matter, when the matter type is a rule adjustment type, all target filling information matched with the protection customization matter is determined from a plurality of preset filling information, and the all target filling information matched with the protection customization matter is filled into the protection customization matter; wherein the plurality of preset filling information includes access control rules, intrusion detection and protection rules, Web application firewall rules and other security rules;

[0187] When the matter type is a configuration parameter adjustment type, a state control instruction and a matter control parameter for the protection customization matter are generated; and the state control instruction and the matter control parameter of the protection customization matter are added to the protection customization matter;

[0188] Wherein, the state control instruction of each protection customization matter is used to control enabling or disabling the protection customization matter; and the matter control parameter of each protection customization matter is used to indicate the execution frequency of the control instruction corresponding to the matter control parameter.

[0189] In the optional embodiment, when the protection scenario is a portal website, all protection sub-items corresponding to the portal website include one or more of DDos attack protection, SQL injection protection, cross-site scripting attack protection, and data backup and recovery;

[0190] When the protection scenario is an e-commerce platform, all protection sub-items corresponding to the portal website include one or more of payment security, user information protection, fraud detection, and business failure monitoring and processing;

[0191] When the protection scenario is the financial industry, all protection sub-items corresponding to the portal website include one or more of data encryption and transmission security, identity authentication and access control, and business continuity monitoring and failure handling;

[0192] When the protection scenario is the gaming industry, all protection sub-items corresponding to the portal website include one or more of DDos attack protection, game data security, anti-cheat mechanism, and server load balancing.

[0193] It can be seen that in the optional embodiment, the item filling mechanism for protection customization items based on different item types is set, wherein for the rule adjustment type, the target filling information matched with the specific protection customization item is determined from a plurality of preset filling information and filled into the item, that is, the corresponding security rules can be flexibly adjusted and applied according to different security requirements and environments, and this mechanism ensures that the security protection measures can accurately target specific threats or vulnerabilities, improving the pertinence and effectiveness of security protection; for the configuration parameter adjustment type, the state control instruction and item control parameter for the specific protection customization item are generated and added to the item, the system allows the user to enable or disable certain protection customization items as needed and set the corresponding execution frequency, so that the user can flexibly and dynamically adjust the security protection strategy according to the actual security situation and resource allocation to optimize performance and resource utilization, that is, it is beneficial to improve the use flexibility and reliability of the subsequent generated customized second protection scheme.

[0194] Embodiment Four

[0195] Please refer to Figure 4 , Figure 4 is another structure diagram of the personalized service intelligent customization device based on the LLM security protection system disclosed in the embodiments of the present application. As shown in Figure 4 , the personalized service intelligent customization device based on the LLM security protection system can include:

[0196] a memory 401 storing executable program codes;

[0197] a processor 402 coupled with the memory 401;

[0198] The processor 402 invokes the executable program code stored in the memory 401 to perform the steps in the personalized service intelligent customization method based on the LLM security protection system described in embodiment one or embodiment two of the present application.

[0199] Embodiment five

[0200] The embodiment of the present application discloses a computer storage medium, which stores computer instructions, and the computer instructions are used to perform the steps in the personalized service intelligent customization method based on the LLM security protection system described in embodiment one or embodiment two of the present application when invoked.

[0201] Embodiment six

[0202] The embodiment of the present application discloses a computer program product, which includes a non-transitory computer storage medium storing a computer program, and the computer program is operable to cause a computer to perform the steps in the personalized service intelligent customization method based on the LLM security protection system described in embodiment one or embodiment two.

[0203] The device embodiments described above are only schematic, and the modules illustrated as separate components can or can not be physically separate, and the components illustrated as modules can or can not be physical modules, i.e., can be located in one place, or can be distributed to multiple network modules. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiment scheme. Those skilled in the art can understand and implement without creative labor.

[0204] Through the specific description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be realized by means of software and the necessary general hardware platform, and of course, can also be realized by hardware. Based on such understanding, the above technical solutions, essentially or in terms of contribution to the prior art, can be embodied in the form of a software product. The computer software product can be stored in a computer storage medium, including a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disk storage, a magnetic disk storage, a magnetic tape storage, or any other computer readable medium that can be used to carry or store data.

[0205] Finally, it should be noted that: the individualized service intelligent customization method and device based on the LLM security protection system disclosed by the embodiments of the present application are only the preferred embodiments of the present application, and are used to illustrate the technical solutions of the present application, but not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that; it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for intelligent customization of personalized services based on an LLM security protection system, characterized by, The method comprises: detecting target user's customization requirement information for the LLM security protection system, and analyzing the customization requirement information to obtain protection scenario information and target requirement details for the LLM security protection system; obtaining a plurality of historical protection schemes recorded in a database and matching the protection scenario information, each of the historical protection schemes comprising a plurality of protection sub-items, each of the protection sub-items being used to adapt to a requirement detail of a historical user; performing requirement detail matching on all the historical protection schemes according to the target requirement details to obtain a first protection scheme with a highest target adaptation value corresponding to the target requirement details; performing scheme customization on the first protection scheme according to the protection scenario information and the target requirement details to obtain a second protection scheme corresponding to the first protection scheme, and performing customization operation on the LLM security protection system according to the second protection scheme; wherein the scheme customization comprises scheme detail adjustment and / or scheme verification; the target requirement details comprise a plurality of requirement customization items matching the protection scenario; the target requirement details further comprise an item type corresponding to each of the requirement customization items, the item type comprising a rule adjustment type or a configuration parameter adjustment type; the scheme customization on the first protection scheme according to the protection scenario information and the target requirement details to obtain the second protection scheme corresponding to the first protection scheme comprises: performing comparison on all the protection sub-items included in the first protection scheme according to all the requirement customization items corresponding to the target requirement details to obtain corresponding similar items, first difference items and second difference items, the first difference items being items remaining after all the requirement customization items are screened from the similar items; the second difference items being items remaining after all the protection sub-items are screened from the similar items; for the first protection scheme, adding the first difference items, retaining the similar items and deleting the second difference items to obtain a new first protection scheme, and determining that the first difference items and the similar items are all protection customization items of the first protection scheme; for each of the protection customization items, performing item filling according to the item type corresponding to the protection customization item to update the protection customization item; after determining that all the protection customization items are updated, determining the new first protection scheme as a second protection scheme; the item filling for each of the protection customization items according to the item type corresponding to the protection customization item comprises: for each of the protection customization items, when the item type is the rule adjustment type, determining all target filling information matching the protection customization item from a plurality of preset filling information, and filling all the target filling information matching the protection customization item into the protection customization item; wherein the plurality of preset filling information comprises access control rules, intrusion detection and protection rules, Web application firewall rules and other security rules. When the matter type is the configuration parameter adjustment type, a state control instruction and a matter control parameter of the protection customization matter are generated, and the state control instruction and the matter control parameter of the protection customization matter are added to the protection customization matter; The state control instruction of each protection customization matter is used to control enabling or disabling the protection customization matter, and the matter control parameter of each protection customization matter is used to indicate the execution frequency of the control instruction corresponding to the matter control parameter. 2.The method of claim 1, wherein, The protection scenario information includes a protection scenario to which the customization requirement information is actually applied, and the protection scenario includes any one of a portal website, an e-commerce platform, a financial industry, and a game industry. The method further includes: obtaining a plurality of historical protection schemes recorded in the database and matching the protection scenario information, including: All basic protection schemes recorded in the database are obtained, and first-level screening is performed on all the basic protection schemes according to a scenario label of the protection scenario in the protection scenario information, to obtain a plurality of first-level screening protection schemes matching the scenario label. Second-level screening is performed on all the first-level screening protection schemes according to a preset screening parameter, to obtain a plurality of second-level screening protection schemes as historical protection schemes; the screening parameter includes a scheme recording date and a scheme quantity; the scheme recording date is used to screen schemes within a target recording date from all the first-level screening protection schemes; and the scheme quantity is used to limit the quantity of all the second-level screening protection schemes. 3.The method of claim 1 or 2, wherein, The method further includes: performing requirement detail pairing on all the historical protection schemes according to the target requirement details, to obtain a first protection scheme with a highest target adaptation value to the target requirement details, including: For each historical protection scheme, all requirement customization matters corresponding to the target requirement details are matched with all protection sub-items corresponding to the historical protection scheme, to obtain first adaptation information of the target requirement details and the historical protection scheme, the first adaptation information including a first adaptation value; According to the customization requirement information, a user group classification of the target user and a corresponding user portrait are determined, and customization preference information adapted to the user group classification is extracted from the user portrait; According to the customization preference information, a scheme adaptation value of each historical protection scheme is evaluated, to obtain second adaptation information of the target user and each historical protection scheme, the second adaptation information including a second adaptation value; According to the first adaptation information and the second adaptation information of each historical protection scheme, a first protection scheme with a highest target adaptation value to the target requirement details is determined. 4.The method of claim 3, wherein, The first adaptation information of each historical protection scheme further includes adaptation matters and difference matters between all protection sub-items corresponding to the historical protection scheme and all requirement customization matters corresponding to the target requirement details. The method further includes: determining, according to the first adaptation information and the second adaptation information of each historical protection scheme, a first protection scheme with a highest target adaptation value to the target requirement details, including: For each of the historical protection schemes, a target adaptation value corresponding to the historical protection scheme and the customized demand information is calculated according to the first adaptation value corresponding to the historical protection scheme and the first preset weight coefficient corresponding to the first adaptation value, the second adaptation value and the second preset weight coefficient corresponding to the second adaptation value; According to the adaptation matters corresponding to the historical protection scheme and the difference matters, adaptation details information of the historical protection scheme is generated; and the adaptation details information is added to the historical protection scheme; A first protection scheme with the highest target adaptation value is selected from all the historical protection schemes. 5.The method of claim 2, wherein, When the protection scene is the e-commerce platform, all the protection sub-items corresponding to the portal website include one or more of payment security, user information protection, anti-fraud detection, and business failure monitoring and processing. When the protection scene is the financial industry, all the protection sub-items corresponding to the portal website include one or more of data encryption and transmission security, identity authentication and access control, and business continuity monitoring and failure handling. When the protection scene is the game industry, all the protection sub-items corresponding to the portal website include one or more of DDos attack protection, game data security, anti-cheating mechanism, and server load balancing. The device comprises:

6. An apparatus for intelligent customization of personalized services based on an LLM security shield system, characterized by, A detection module for detecting customized demand information of a target user for an LLM security protection system; A demand analysis module for analyzing the customized demand information to obtain protection scene information and target demand details for the LLM security protection system; An acquisition module for acquiring a plurality of historical protection schemes recorded in a database and matching the protection scene information, each of the historical protection schemes comprising a plurality of protection sub-items, each of the protection sub-items being used to adapt to a demand detail of a historical user; A pairing module for performing demand detail pairing on all the historical protection schemes according to the target demand details to obtain a first protection scheme with the highest target adaptation value corresponding to the target demand details; A scheme customization module for performing scheme customization on the first protection scheme according to the protection scene information and the target demand details to obtain a second protection scheme corresponding to the first protection scheme, and performing customization operation on the LLM security protection system according to the second protection scheme; wherein the scheme customization comprises scheme detail adjustment and / or scheme verification; The target demand details comprise a plurality of demand customization matters matching the protection scene; the target demand details further comprise a matter type corresponding to each of the demand customization matters, the matter type comprising a rule adjustment type or a configuration parameter adjustment type; ​ The scheme customization module performs scheme customization on the first protection scheme according to the protection scenario information and the target requirement details, and obtains a second protection scheme corresponding to the first protection scheme. All the requirement customization matters corresponding to the target requirement details are compared with all the protection sub-items included in the first protection scheme, and corresponding similar matters, first difference matters and second difference matters are obtained. The first difference matters are the matters remaining after all the requirement customization matters are filtered from the similar matters. The second difference matters are the matters remaining after all the protection sub-items are filtered from the similar matters. For the first protection scheme, the first difference matters are added, the similar matters are retained, and the second difference matters are deleted, to obtain a new first protection scheme. Meanwhile, the first difference matters and the similar matters are determined as all the protection customization matters of the first protection scheme. For each protection customization matter, matter filling is performed according to the matter type corresponding to the protection customization matter, to update the protection customization matter. After determining that all the protection customization matters are updated, the new first protection scheme is determined as a second protection scheme. The scheme customization module performs matter filling for each protection customization matter according to the matter type corresponding to the protection customization matter. For each protection customization matter, when the matter type is the rule adjustment type, all target filling information matched with the protection customization matter is determined from a plurality of preset filling information, and the target filling information matched with the protection customization matter is filled into the protection customization matter. The plurality of preset filling information includes access control rules, intrusion detection and protection rules, Web application firewall rules, and other security rules. When the matter type is the configuration parameter adjustment type, a state control instruction and a matter control parameter for the protection customization matter are generated, and the state control instruction and the matter control parameter of the protection customization matter are added to the protection customization matter. The state control instruction of each protection customization matter is used to control enabling or disabling the protection customization matter. The matter control parameter of each protection customization matter is used to indicate the execution frequency of the control instruction corresponding to the matter control parameter.

7. An apparatus for intelligent customization of personalized services based on an LLM security shield system, characterized by, The apparatus includes: a memory storing executable program codes; a processor coupled with the memory; the processor invokes the executable program codes stored in the memory to execute the method for intelligent customization of individual services of an LLM-based security protection system according to any one of claims 1-5.

8. A computer storage medium, characterized in that, The computer storage medium stores computer instructions, which are invoked to execute the method for intelligent customization of individual services of an LLM-based security protection system according to any one of claims 1-5.

Images