A knowledge base cloud sharing method for enterprises based on AI big model

Abstract

A method for sharing knowledge base cloud for enterprises based on AI big model relates to the technical field of data sharing. Blockchain technology is used to build an enterprise knowledge base sharing platform, and the enterprise knowledge base sharing platform is communicatively connected with a plurality of user interaction terminals. When the user interaction terminal interacts with enterprise data, a flow prediction model and an intrusion interaction prediction model are constructed, and flow tracking analysis is performed on the user interaction terminal. The adaptive interaction level of the user interaction terminal and the adaptive security level of enterprise data are set according to the analysis results, and the user interaction process is controlled according to the adaptive interaction level and the adaptive security level. The end timestamp of the unit cycle of the user interaction terminal is used as a detection time point, and intrusion security detection is performed on the user interaction terminal at the detection time point. The intrusion interaction prediction model and the flow prediction model are trained in real time according to the intrusion detection result to ensure the efficiency and security of enterprise data knowledge sharing.

Description

Technical Field

[0001] The present invention relates to the technical field of data sharing, and in particular to an enterprise knowledge base cloud sharing method based on an AI big model. Background Art

[0002] The prior art CN116756085A "An intelligent collaborative control and management system for enterprise knowledge base" includes: an upload management module, used to manage file upload operations on the enterprise knowledge base; a file management module, used to manage files in the enterprise knowledge base; a user management module, used to manage users within the enterprise; a device management module, used to manage the terminal sharing devices of the enterprise; a file sharing module, used to manage file sharing, wherein the file sharing module includes a first sharing unit and a second sharing unit, based on WeChat, which is a standard feature of everyone, as a carrier, users can conveniently share the enterprise knowledge base content of the enterprise group scale through basic mobile terminals, and set differentiated requirements for files in the enterprise knowledge base according to the organizational price of the enterprise, so as to realize highly collaborative knowledge base content management and improve the efficiency of knowledge sharing of the enterprise.

[0003] Prior art CN114625797A "A method, device and electronic device for sharing a knowledge base" The method includes: obtaining knowledge requirement information transmitted by a second node from a blockchain; determining a first knowledge content based on the knowledge requirement information; transmitting the first knowledge content to the blockchain; wherein the first node and the second node share the knowledge content in the blockchain, the second node transmits the knowledge requirement information to the blockchain, the first node determines the specific knowledge content based on the knowledge requirement information in the blockchain, and transmits the knowledge content to the blockchain, and the blockchain is used as a knowledge base to realize the sharing of knowledge content; In an embodiment of the present application, the first node and the second node jointly participate in the construction of a knowledge base based on a blockchain, which can realize the secure sharing of the knowledge base between all parties and realize the unified management of the knowledge base.

[0004] The existing blockchain evidence system can only prove the authenticity and immutability of the enterprise data shared or accessed by the user interaction terminal in the blockchain network. It cannot ensure whether there are problems with the enterprise data shared by the user interaction terminal and the user interaction terminal itself. In addition, the enterprise knowledge base network data is of various types and huge in volume. Due to the complexity of the network structure, there are more and more attacks and malicious behaviors against the network, and attacks on the network system using vulnerabilities and security defects in the network are emerging in an endless stream. However, most of the current intrusion detection algorithms stop at labeling each piece of data as normal or intrusion. Some of these intrusions are false intrusions. A large amount of normal enterprise data that is misjudged as intrusion is mixed with a small amount of abnormal data that is correctly judged as abnormal, which will seriously weaken the availability of the current intrusion detection algorithm. The anomaly detection algorithm often only finds anomalies based on data, and a high proportion of false anomalies often affects the subsequent abnormal behavior association. When the enterprise knowledge base network encounters an attack, a large amount of abnormal data is generated. It is difficult to explore the laws of a large amount of data by manual means. How to perceive and predict security events, display the current network security risks in real time, and predict possible security events are problems that we urgently need to solve. Summary of the invention

[0005] In order to solve the above technical problems, the purpose of the present invention is to provide an enterprise knowledge base cloud sharing method based on an AI big model, comprising the following steps:

[0006] Step s1: constructing an enterprise knowledge base sharing platform using blockchain technology, wherein a plurality of blockchain nodes are arranged in the enterprise knowledge base sharing platform, wherein the blockchain nodes are used to store enterprise data of the enterprise knowledge base, and the plurality of blockchain nodes are linked to each other to form a blockchain network, and the blockchain nodes are communicatively connected to a plurality of user interaction terminals;

[0007] Step s2: When the user interaction terminal interacts with enterprise data, a power flow prediction model and an intrusion interaction prediction model are constructed, and a power flow tracking analysis is performed on the user interaction terminal. The adaptive interaction level of the user interaction terminal and the adaptive security level of the enterprise data are set according to the analysis results, and the user interaction process is controlled according to the adaptive interaction level and the adaptive security level;

[0008] Step s3: Taking the end timestamp of the unit cycle of the user interaction end as the detection time point, performing intrusion security detection on the user interaction end at the detection time point, and training the intrusion interaction prediction model and the power flow prediction model in real time according to the intrusion detection results.

[0009] Furthermore, the process of sharing enterprise information on the user interaction end includes:

[0010] When the user interaction end interacts with enterprise data, a trend tracking analysis is performed on the user interaction end, and the adaptive interaction level of the user interaction end is set according to the analysis results. If the user interaction end shares enterprise data with the blockchain node, a sharing level threshold is set. If the adaptive interaction level is greater than or equal to the sharing level threshold, the enterprise data is encrypted and uploaded to the blockchain node. The blockchain node decrypts the received enterprise data and packages it into a new block, broadcasts the new block to the entire blockchain network, adds the new block to the end of the blockchain, and updates the enterprise data in the new block to the enterprise knowledge base of all blockchain nodes. If the adaptive interaction level is less than the sharing level threshold, the user interaction end is authenticated. If the user interaction end passes the authenticity verification, the enterprise data is encrypted and uploaded to the blockchain node. If the user interaction end fails the authenticity verification, the user interaction end is marked as a sudden interaction intrusion.

[0011] Furthermore, the process of the user interaction end accessing enterprise information includes:

[0012] When the user interaction terminal accesses the enterprise data of the blockchain node, a flow tracking analysis is performed on the enterprise data of the user interaction terminal and the blockchain node, and the adaptive interaction level of the user interaction terminal and the adaptive security level of the accessed enterprise data are set according to the analysis results. If the adaptive interaction level is greater than or equal to the adaptive security level, the authenticity of the user interaction terminal is verified. If the user interaction terminal passes the authenticity verification, the enterprise data is transmitted to the user interaction terminal. If the user interaction terminal fails the authenticity verification, the user interaction terminal is marked as a sudden interaction intrusion;

[0013] If the adaptive interaction level is lower than the adaptive security level, the user interaction terminal is instructed to submit a manual review application document to the enterprise knowledge base sharing platform, and the administrator of the enterprise knowledge base sharing platform determines whether to grant access rights based on the manual review application document.

[0014] Furthermore, the process of constructing a power flow prediction model and an intrusion interaction prediction model and obtaining normal interaction prediction information and intrusion interaction prediction information of each user interaction terminal includes:

[0015] Construct a power flow prediction model based on RBF neural network, obtain historical normal interaction records between all user interaction terminals and blockchain nodes, the historical normal interaction records include interaction time, enterprise data type and enterprise data size within a unit period, obtain training data through the historical normal interaction records, use the training data to train the power flow prediction model, and output interaction prediction information according to the trained power flow prediction model, the normal interaction prediction information includes data interaction probability of enterprise data type and enterprise data size at each interaction time within a unit period of all user interaction terminals, and the data interaction probability includes data sharing probability and data access probability;

[0016] At the same time, an intrusion interaction prediction model is constructed based on the RBF neural network to obtain historical intrusion interaction data of sudden interaction intrusions at all user interaction terminals, wherein the historical intrusion interaction records include intrusion interaction moments, intrusion enterprise data types, and intrusion enterprise data sizes within a unit period, and training data is obtained through the historical intrusion interaction records. The intrusion interaction prediction model is trained using the training data, and intrusion interaction prediction information is output according to the trained intrusion interaction prediction model. The intrusion interaction prediction information includes the intrusion data interaction probability of the intrusion enterprise data type and the intrusion enterprise data size at each intrusion interaction moment within a unit period of all user interaction terminals, and the intrusion data interaction probability includes the intrusion data sharing probability and the intrusion data access probability.

[0017] Furthermore, the process of obtaining the adaptive security level corresponding to the enterprise data type and enterprise data size at a plurality of interaction moments includes:

[0018] The normal interaction prediction information is associated and aggregated with the intrusion interaction prediction information to obtain the similarity between the enterprise data type and enterprise data size at each interaction moment and the intrusion enterprise data type and intrusion enterprise data size at each intrusion interaction moment, preset a similarity threshold, obtain the enterprise data type and enterprise data size at several interaction moments with a similarity greater than the similarity threshold and the intrusion enterprise data type and intrusion enterprise data size at the corresponding intrusion interaction moment, and obtain the adaptive security level corresponding to the enterprise data type and enterprise data size at the several interaction moments based on the data interaction probabilities of the enterprise data types and enterprise data sizes at the several interaction moments and the intrusion data interaction probabilities of the intrusion enterprise data types and intrusion enterprise data sizes at the corresponding intrusion interaction moments.

[0019] Furthermore, the process of obtaining the adaptive interaction level of the user interaction terminal includes:

[0020] The enterprise data type and enterprise data size at the current interaction moment of the user interaction terminal are obtained, and after normalizing the data interaction probabilities of the enterprise data types and enterprise data sizes at each interaction moment within a unit period of the user interaction terminal, the normalized data interaction probability coefficients of the enterprise data types and enterprise data sizes at the current interaction moment of the user interaction terminal are obtained, and an adaptive interaction level comparison table is pre-set, wherein the adaptive interaction level comparison table includes adaptive interaction level comparison tables corresponding to different normalized data interaction probability coefficient threshold intervals, and the adaptive interaction level of the user interaction terminal is obtained according to the normalized data interaction probability coefficient and the adaptive interaction level comparison table.

[0021] Furthermore, the end timestamp of the unit cycle of the user interaction end is used as the detection time point, and the user interaction end is subjected to intrusion security detection at the detection time point. The process of real-time training of the intrusion interaction prediction model and the power flow prediction model according to the intrusion detection result includes:

[0022] The end timestamp of the unit cycle of the user interaction terminal is used as the detection time point, and the user interaction terminal is subjected to intrusion security detection at the detection time point. If an intrusion interaction record is detected in the user interaction terminal and the enterprise knowledge base sharing platform does not mark the sudden interaction intrusion of the user interaction terminal, the intrusion interaction record is extracted as the training data of the intrusion interaction prediction model, and the intrusion interaction prediction model is retrained;

[0023] At the same time, normal interaction records within a unit period of the user interaction terminal are obtained, and the normal interaction records are used as training data of the power flow prediction model to retrain the power flow prediction model.

[0024] Furthermore, the process of verifying the authenticity of the user interaction terminal includes:

[0025] The enterprise knowledge base sharing platform pre-sets the biometric database of each user interaction end, obtains the biometric data of the user interaction end to be authenticated, and matches the biometric data with the biometric data of the user interaction end in the biometric database for consistency. If the match is successful, the user interaction end passes the authenticity verification; if the match fails, the user interaction end fails the authenticity verification.

[0026] Compared with the prior art, the present invention has the following beneficial effects:

[0027] 1. Use blockchain technology to build an enterprise knowledge base sharing platform, in which a number of blockchain nodes are set up, and the blockchain nodes are used to store enterprise data of the enterprise knowledge base, and the number of blockchain nodes are linked to each other to form a blockchain network, and the blockchain nodes are connected to a number of user interaction terminals; when the user interaction terminal interacts with enterprise data, a flow prediction model and an intrusion interaction prediction model are constructed, and a flow tracking analysis is performed on the user interaction terminal. According to the analysis results, the adaptive interaction level of the user interaction terminal and the adaptive security level of the enterprise data are set, and the user interaction process is controlled according to the adaptive interaction level and the adaptive security level, and personalized access control is performed on different user interaction terminals to ensure the efficiency and security of enterprise data knowledge sharing;

[0028] 2. Take the end timestamp of the unit cycle of the user interaction end as the detection time point, perform intrusion security detection on the user interaction end at the detection time point, and train the intrusion interaction prediction model and the flow prediction model in real time according to the intrusion detection results, display the current network security risks and predict possible security incidents in real time, and ensure the accuracy of the intrusion interaction prediction model and the flow prediction model in real time. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 This is a schematic diagram of a method for sharing enterprise knowledge base cloud based on an AI big model according to an embodiment of the present application. DETAILED DESCRIPTION

[0030] The following is a clear and complete description of the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0031] like Figure 1 As shown, a method for sharing enterprise knowledge base cloud based on AI big model includes the following steps:

[0032] Step s1: constructing an enterprise knowledge base sharing platform using blockchain technology, wherein a plurality of blockchain nodes are arranged in the enterprise knowledge base sharing platform, wherein the blockchain nodes are used to store enterprise data of the enterprise knowledge base, and the plurality of blockchain nodes are linked to each other to form a blockchain network, and the blockchain nodes are communicatively connected to a plurality of user interaction terminals;

[0033] Step s2: When the user interaction terminal interacts with enterprise data, a power flow prediction model and an intrusion interaction prediction model are constructed, and a power flow tracking analysis is performed on the user interaction terminal. The adaptive interaction level of the user interaction terminal and the adaptive security level of the enterprise data are set according to the analysis results, and the user interaction process is controlled according to the adaptive interaction level and the adaptive security level;

[0034] Step s3: Taking the end timestamp of the unit cycle of the user interaction end as the detection time point, performing intrusion security detection on the user interaction end at the detection time point, and training the intrusion interaction prediction model and the power flow prediction model in real time according to the intrusion detection results.

[0035] It should be further explained that, in the specific implementation process, the process of sharing enterprise information on the user interaction end includes:

[0036] When the user interaction end interacts with enterprise data, a trend tracking analysis is performed on the user interaction end, and the adaptive interaction level of the user interaction end is set according to the analysis results. If the user interaction end shares enterprise data with the blockchain node, a sharing level threshold is set. If the adaptive interaction level is greater than or equal to the sharing level threshold, the enterprise data is encrypted and uploaded to the blockchain node. The blockchain node decrypts the received enterprise data and packages it into a new block, broadcasts the new block to the entire blockchain network, adds the new block to the end of the blockchain, and updates the enterprise data in the new block to the enterprise knowledge base of all blockchain nodes. If the adaptive interaction level is less than the sharing level threshold, the user interaction end is authenticated. If the user interaction end passes the authenticity verification, the enterprise data is encrypted and uploaded to the blockchain node. If the user interaction end fails the authenticity verification, the user interaction end is marked as a sudden interaction intrusion.

[0037] It should be further explained that, in the specific implementation process, the process of the user interaction end accessing the enterprise information includes:

[0038] When the user interaction terminal accesses the enterprise data of the blockchain node, a flow tracking analysis is performed on the enterprise data of the user interaction terminal and the blockchain node, and the adaptive interaction level of the user interaction terminal and the adaptive security level of the accessed enterprise data are set according to the analysis results. If the adaptive interaction level is greater than or equal to the adaptive security level, the authenticity of the user interaction terminal is verified. If the user interaction terminal passes the authenticity verification, the enterprise data is transmitted to the user interaction terminal. If the user interaction terminal fails the authenticity verification, the user interaction terminal is marked as a sudden interaction intrusion;

[0039] If the adaptive interaction level is lower than the adaptive security level, the user interaction terminal is instructed to submit a manual review application document to the enterprise knowledge base sharing platform, and the administrator of the enterprise knowledge base sharing platform determines whether to grant access rights based on the manual review application document.

[0040] It should be further explained that, in the specific implementation process, the process of constructing the power flow prediction model and the intrusion interaction prediction model and obtaining the normal interaction prediction information and the intrusion interaction prediction information of each user interaction terminal includes:

[0041] A power flow prediction model is constructed based on an RBF neural network to obtain historical normal interaction records between all user interaction terminals and blockchain nodes. The historical normal interaction records include the interaction time, enterprise data type, and enterprise data size within a unit period. Training data is obtained through the historical normal interaction records, and the power flow prediction model is trained using the training data. Interaction prediction information is output according to the trained power flow prediction model. The unit period is 1 day. The normal interaction prediction information includes the data interaction probability of the enterprise data type and enterprise data size at each interaction time within the unit period of all user interaction terminals. The data interaction probability includes the data sharing probability and the data access probability.

[0042] At the same time, an intrusion interaction prediction model is constructed based on the RBF neural network to obtain historical intrusion interaction data of sudden interaction intrusions at all user interaction terminals. The interaction intrusion includes sharing false enterprise data to the enterprise knowledge base or accessing and stealing enterprise data from the enterprise knowledge base after the user interaction terminal is invaded by a hacker. The historical intrusion interaction record includes the intrusion interaction moment, the intrusion enterprise data type and the intrusion enterprise data size within a unit period. Training data is obtained through the historical intrusion interaction record, and the intrusion interaction prediction model is trained using the training data. Intrusion interaction prediction information is output according to the trained intrusion interaction prediction model. The intrusion interaction prediction information includes the intrusion data interaction probability of the intrusion enterprise data type and the intrusion enterprise data size at each intrusion interaction moment within the unit period of all user interaction terminals, and the intrusion data interaction probability includes the intrusion data sharing probability and the intrusion data access probability.

[0043] It should be further explained that, in a specific implementation process, the process of obtaining the adaptive security level corresponding to the enterprise data type and enterprise data size at a number of interaction moments includes:

[0044] The normal interaction prediction information is associated and aggregated with the intrusion interaction prediction information to obtain the similarity between the enterprise data type and enterprise data size at each interaction moment and the intrusion enterprise data type and intrusion enterprise data size at each intrusion interaction moment, preset a similarity threshold, obtain the enterprise data type and enterprise data size at several interaction moments with a similarity greater than the similarity threshold and the intrusion enterprise data type and intrusion enterprise data size at the corresponding intrusion interaction moment, and obtain the adaptive security level corresponding to the enterprise data type and enterprise data size at the several interaction moments based on the data interaction probabilities of the enterprise data types and enterprise data sizes at the several interaction moments and the intrusion data interaction probabilities of the intrusion enterprise data types and intrusion enterprise data sizes at the corresponding intrusion interaction moments.

[0045] It should be further explained that, in the specific implementation process, the calculation formula for obtaining the similarity between the enterprise data type and enterprise data size at each interaction moment and the intrusion enterprise data type and intrusion enterprise data size at each intrusion interaction moment is:

[0046] ;

[0047] ;

[0048] in, represents the similarity between normal interaction prediction information i and intrusion interaction prediction information j, represents the enterprise data size of normal interactive prediction information i, represents the enterprise data size of intrusion interaction prediction information j, Normal interaction prediction information i’s interaction time, represents the interaction time of the intrusion interaction prediction information j, The enterprise data type representing normal interaction prediction information i, The enterprise data type of intrusion interaction prediction information j, , represents the weight factor;

[0049] It should be further explained that, in the specific implementation process, the calculation formula for obtaining the adaptive security level corresponding to the enterprise data type and enterprise data size at several interaction moments according to the data interaction probability of the enterprise data type and enterprise data size at the several interaction moments and the intrusion data interaction probability of the intrusion enterprise data type and intrusion enterprise data size at the corresponding intrusion interaction moments is:

[0050] ;

[0051] Among them, sty represents the adaptive security level, LP represents the data interaction probability of enterprise data type and enterprise data size at several interaction moments, LD represents the intrusion data interaction probability of intrusion enterprise data type and intrusion enterprise data size at the intrusion interaction moment, , Indicates the conversion factor.

[0052] It should be further explained that, in a specific implementation process, the process of obtaining the adaptive interaction level of the user interaction terminal includes:

[0053] The enterprise data type and enterprise data size at the current interaction moment of the user interaction terminal are obtained, and after normalizing the data interaction probabilities of the enterprise data types and enterprise data sizes at each interaction moment within a unit period of the user interaction terminal, the normalized data interaction probability coefficients of the enterprise data types and enterprise data sizes at the current interaction moment of the user interaction terminal are obtained, and an adaptive interaction level comparison table is pre-set, wherein the adaptive interaction level comparison table includes adaptive interaction level comparison tables corresponding to different normalized data interaction probability coefficient threshold intervals, and the adaptive interaction level of the user interaction terminal is obtained according to the normalized data interaction probability coefficient and the adaptive interaction level comparison table.

[0054] It should be further explained that, in the specific implementation process, the calculation formula for obtaining the normalized data interaction probability coefficient of the enterprise data type and enterprise data size at the current interaction moment of the user interaction terminal is:

[0055] ;

[0056] in, The normalized data interaction probability coefficient representing the enterprise data type and enterprise data size at the current interaction moment of the user interaction terminal, The data interaction probability representing the enterprise data type and enterprise data size at the current interaction moment of the user interaction end, The maximum data interaction probability of the enterprise data type and enterprise data size at each interaction moment within the unit cycle of the user interaction end, The minimum data interaction probability of the enterprise data type and enterprise data size at each interaction moment within a unit cycle of the user interaction end.

[0057] It should be further explained that, in the specific implementation process, the end timestamp of the unit cycle of the user interaction end is used as the detection time point, and the intrusion security detection is performed on the user interaction end at the detection time point. The process of real-time training of the intrusion interaction prediction model and the power flow prediction model according to the intrusion detection result includes:

[0058] The end timestamp of the unit cycle of the user interaction end is used as the detection time point, and the user interaction end is subjected to intrusion security detection at the detection time point. For example, vulnerability scanning software or anti-virus software is used to scan the software or system vulnerabilities of the user interaction end, and known vulnerabilities are patched or viruses are detected in a timely manner. If an intrusion interaction record is detected in the user interaction end and the enterprise knowledge base sharing platform does not mark the sudden interaction intrusion of the user interaction end, the intrusion interaction record is extracted as the training data of the intrusion interaction prediction model, and the intrusion interaction prediction model is retrained;

[0059] At the same time, normal interaction records within a unit period of the user interaction terminal are obtained, and the normal interaction records are used as training data of the power flow prediction model to retrain the power flow prediction model.

[0060] It should be further explained that, in the specific implementation process, the process of verifying the authenticity of the user interaction terminal includes:

[0061] The enterprise knowledge base sharing platform pre-sets the biometric database of each user interaction terminal, obtains the biometric data of the user interaction terminal to be verified for authenticity, the biometric data includes fingerprint data, face data, iris data, etc., and matches the biometric data with the biometric data of the user interaction terminal in the biometric database for consistency. If the match is successful, the user interaction terminal passes the authenticity verification. If the match fails, the user interaction terminal fails to pass the authenticity verification.

[0062] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A knowledge base cloud sharing method for enterprises based on AI big model, characterized in that: The following steps are involved: Step s1: constructing an enterprise knowledge base sharing platform using blockchain technology, wherein a plurality of blockchain nodes are arranged in the enterprise knowledge base sharing platform, wherein the blockchain nodes are used to store enterprise data of the enterprise knowledge base, and the plurality of blockchain nodes are linked to each other to form a blockchain network, and the blockchain nodes are communicatively connected to a plurality of user interaction terminals; Step s2: When the user interaction end interacts with enterprise data, a power flow prediction model and an intrusion interaction prediction model are constructed. If the user interaction end shares enterprise data with the blockchain node, a power flow tracking analysis is performed on the user interaction end. The adaptive interaction level of the user interaction end is set according to the analysis result, and a sharing level threshold is set. If the adaptive interaction level is greater than or equal to the sharing level threshold, the enterprise data is encrypted and uploaded to the blockchain node. The blockchain node decrypts the received enterprise data and packages it into a new block, broadcasts the new block to the entire blockchain network, adds the new block to the end of the blockchain, and updates the enterprise data in the new block to the enterprise knowledge base of all blockchain nodes. If the adaptive interaction level is less than the sharing level threshold, the user interaction end is authenticated. If the user interaction end passes the authenticity verification, the enterprise data is encrypted and uploaded to the blockchain node. If the user interaction end fails the authenticity verification, the user interaction end is marked as a sudden interaction intrusion. When the user interaction terminal accesses the enterprise data of the blockchain node, a flow tracking analysis is performed on the enterprise data of the user interaction terminal and the blockchain node, and the adaptive interaction level of the user interaction terminal and the adaptive security level of the accessed enterprise data are set according to the analysis results. If the adaptive interaction level is greater than or equal to the adaptive security level, the authenticity of the user interaction terminal is verified. If the user interaction terminal passes the authenticity verification, the enterprise data is transmitted to the user interaction terminal. If the user interaction terminal fails the authenticity verification, the user interaction terminal is marked as a sudden interaction intrusion; If the adaptive interaction level is lower than the adaptive security level, the user interaction terminal is ordered to submit a manual review application document to the enterprise knowledge base sharing platform, and the administrator of the enterprise knowledge base sharing platform determines whether to grant access rights based on the manual review application document; Step s3: Taking the end timestamp of the unit cycle of the user interaction end as the detection time point, performing intrusion security detection on the user interaction end at the detection time point, and training the intrusion interaction prediction model and the power flow prediction model in real time according to the intrusion detection results.

2. According to the enterprise knowledge base cloud sharing method based on AI big model according to claim 1, it is characterized in that: The process of constructing a power flow prediction model and an intrusion interaction prediction model and obtaining normal interaction prediction information and intrusion interaction prediction information of each user interaction terminal includes: Construct a power flow prediction model based on RBF neural network, obtain historical normal interaction records between all user interaction terminals and blockchain nodes, the historical normal interaction records include interaction time, enterprise data type and enterprise data size within a unit period, obtain training data through the historical normal interaction records, use the training data to train the power flow prediction model, and output normal interaction prediction information according to the trained power flow prediction model, the normal interaction prediction information includes data interaction probability of enterprise data type and enterprise data size at each interaction time within a unit period of all user interaction terminals, and the data interaction probability includes data sharing probability and data access probability; At the same time, an intrusion interaction prediction model is constructed based on the RBF neural network, and historical intrusion interaction data of sudden interaction intrusions at all user interaction terminals are obtained, the historical intrusion interaction data including the intrusion interaction moment, the intrusion enterprise data type and the intrusion enterprise data size within a unit period, training data is obtained through the historical intrusion interaction data, the intrusion interaction prediction model is trained using the training data, and intrusion interaction prediction information is output according to the trained intrusion interaction prediction model, the intrusion interaction prediction information including the intrusion data interaction probability of the intrusion enterprise data type and the intrusion enterprise data size at each intrusion interaction moment within the unit period of all user interaction terminals, and the intrusion data interaction probability includes the intrusion data sharing probability and the intrusion data access probability.

3. According to claim 2, a method for sharing enterprise knowledge base cloud based on AI big model is characterized in that: The process of obtaining the adaptive security level corresponding to the enterprise data type and enterprise data size at a number of interaction moments includes: The normal interaction prediction information is associated and aggregated with the intrusion interaction prediction information to obtain the similarity between the enterprise data type and enterprise data size at each interaction moment and the intrusion enterprise data type and intrusion enterprise data size at each intrusion interaction moment, preset a similarity threshold, obtain the enterprise data type and enterprise data size at several interaction moments with a similarity greater than the similarity threshold and the intrusion enterprise data type and intrusion enterprise data size at the corresponding intrusion interaction moment, and obtain the adaptive security level corresponding to the enterprise data type and enterprise data size at the several interaction moments based on the data interaction probabilities of the enterprise data types and enterprise data sizes at the several interaction moments and the intrusion data interaction probabilities of the intrusion enterprise data types and intrusion enterprise data sizes at the corresponding intrusion interaction moments.

4. According to claim 3, a method for sharing enterprise knowledge base cloud based on AI big model is characterized in that: The process of obtaining the adaptive interaction level of the user interaction terminal includes: The enterprise data type and enterprise data size at the current interaction moment of the user interaction terminal are obtained, and after normalizing the data interaction probabilities of the enterprise data types and enterprise data sizes at each interaction moment within a unit period of the user interaction terminal, the normalized data interaction probability coefficients of the enterprise data types and enterprise data sizes at the current interaction moment of the user interaction terminal are obtained, and an adaptive interaction level comparison table is pre-set, wherein the adaptive interaction level comparison table includes adaptive interaction level comparison tables corresponding to different normalized data interaction probability coefficient threshold intervals, and the adaptive interaction level of the user interaction terminal is obtained according to the normalized data interaction probability coefficient and the adaptive interaction level comparison table.

5. According to claim 4, a method for sharing enterprise knowledge base cloud based on AI big model is characterized in that: The end timestamp of the unit cycle of the user interaction end is used as the detection time point, and the intrusion security detection is performed on the user interaction end at the detection time point. The process of real-time training of the intrusion interaction prediction model and the power flow prediction model according to the intrusion detection result includes: The end timestamp of the unit cycle of the user interaction terminal is used as the detection time point, and the user interaction terminal is subjected to intrusion security detection at the detection time point. If an intrusion interaction record is detected in the user interaction terminal and the enterprise knowledge base sharing platform does not mark the sudden interaction intrusion of the user interaction terminal, the intrusion interaction record is extracted as the training data of the intrusion interaction prediction model, and the intrusion interaction prediction model is retrained; At the same time, normal interaction records within a unit period of the user interaction terminal are obtained, and the normal interaction records are used as training data of the power flow prediction model to retrain the power flow prediction model.

6. The enterprise knowledge base cloud sharing method based on AI big model according to claim 5 is characterized in that: The process of verifying the authenticity of the user interaction end includes: The enterprise knowledge base sharing platform pre-sets the biometric database of each user interaction end, obtains the biometric data of the user interaction end to be authenticated, and matches the biometric data with the biometric data of the user interaction end in the biometric database for consistency. If the match is successful, the user interaction end passes the authenticity verification; if the match fails, the user interaction end fails the authenticity verification.

Images