CN115118462B - Data privacy protection method based on convolution enhancement chain - Google Patents
Data privacy protection method based on convolution enhancement chain Download PDFInfo
- Publication number
- CN115118462B CN115118462B CN202210646887.6A CN202210646887A CN115118462B CN 115118462 B CN115118462 B CN 115118462B CN 202210646887 A CN202210646887 A CN 202210646887A CN 115118462 B CN115118462 B CN 115118462B
- Authority
- CN
- China
- Prior art keywords
- data
- learning
- privacy protection
- convolution
- blockchain network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L63/00—Network architectures or network communication protocols for network security
- H04L63/04—Network architectures or network communication protocols for network security for providing a confidential data exchange among entities communicating through data packet networks
- H04L63/0428—Network architectures or network communication protocols for network security for providing a confidential data exchange among entities communicating through data packet networks wherein the data content is protected, e.g. by encrypting or encapsulating the payload
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/16—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks using machine learning or artificial intelligence
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L63/00—Network architectures or network communication protocols for network security
- H04L63/20—Network architectures or network communication protocols for network security for managing network security; network security policies in general
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/14—Session management
- H04L67/141—Setup of application sessions
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S40/00—Systems for electrical power generation, transmission, distribution or end-user application management characterised by the use of communication or information technologies, or communication or information technology specific aspects supporting them
- Y04S40/20—Information technology specific aspects, e.g. CAD, simulation, modelling, system security
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Computer Security & Cryptography (AREA)
- Signal Processing (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Hardware Design (AREA)
- Artificial Intelligence (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Databases & Information Systems (AREA)
- Medical Informatics (AREA)
- Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The application discloses a data privacy protection method based on a convolution enhancement chain, which comprises the following steps: acquiring learning data corresponding to each learning subject, establishing corresponding private data nodes according to the learning subjects, and storing the learning data into the corresponding private data nodes; establishing a secure network connection between the private data nodes based on intelligent contracts and consensus protocols of a blockchain network, and mutually communicating to realize self state update; and constructing a bottom layer system based on reinforcement learning and security calculation at each private data node, forming a convolution enhancement chain together with the blockchain network, and carrying out privacy protection on the learning data. The method can solve the problems that the availability of data based on the protection of the traditional data privacy protection mechanism is not high, the model training effect is poor, or the data in the model is completely revealed after the model is attacked.
Description
Technical Field
The present application relates to the technical fields of blockchain networks, information security and artificial intelligence, and more particularly, to a data privacy protection method based on a convolution enhancement chain.
Background
With the development of information technology, the trend of integrating network, physical and social systems into a highly unified information society, not just the digital internet, is becoming more and more evident. In such an information society, data is an asset of its owner, the use of which should be completely controlled by its owner, but there is a problem that a large amount of data is maliciously used, in order to achieve privacy protection of data, from the most primitive encryption mechanism to the privacy protection method based on differential privacy and secure multiparty computation today, to the horizontal blank of the blockchain network, a certain effect is achieved for data privacy protection to a certain extent.
However, as time accumulates and technology innovates, the protection effect and degree of the technology is greatly compromised, which also forces the privacy protection technology to advance. In recent years, the achievement and development of the artificial intelligence field are witnessed, and the combination of artificial intelligence and privacy protection technology is continuously excavated and developed. Among them, the combination of the fields of machine learning and security computing is particularly close, and representative examples are federal learning and edge computing. By combining a data source with a model algorithm of machine learning or deep learning, the models are trained by utilizing huge data, a strong pre-trained model is obtained, and then a privacy protection mechanism is intelligently executed. However, the usability of the data after privacy protection is not high, so that the model training effect is poor or the data in the model is totally revealed after the model is attacked.
Disclosure of Invention
Aiming at least one defect or improvement requirement of the prior art, the invention provides a data privacy protection method, a device, electronic equipment and a computer readable storage medium based on a convolution enhancement chain, which aim to solve the problems that the usability of data protected based on a traditional data privacy protection mechanism is not high, the model training effect is poor or the data in the model is totally revealed after the model is attacked.
To achieve the above object, according to a first aspect of the present invention, there is provided a data privacy protection method based on a convolution enhancement chain, including: acquiring learning data corresponding to each learning subject, establishing corresponding private data nodes according to the learning subjects, and storing the learning data into the corresponding private data nodes; establishing a secure network connection between the private data nodes based on intelligent contracts and consensus protocols of a blockchain network, and mutually communicating to realize self state update; and constructing a bottom layer system based on reinforcement learning and security calculation at each private data node, forming a convolution enhancement chain together with the blockchain network, and carrying out privacy protection on the learning data.
In one embodiment of the present invention, the establishing a secure network connection between the private data nodes and communicating with each other to implement self-status update includes: a process control plate and a data storage center are respectively arranged in the private data node to respectively control two processes of a data input node and a data output node; and constructing a blockchain network bridge between the process control plate and the data storage center, wherein the blockchain network bridge is used for receiving response information of other private data nodes in real time and updating self state information according to the response information.
In one embodiment of the present invention, said constructing an underlying system based on reinforcement learning and security computation at each of said private data nodes, together with said blockchain network, forms a convolutional reinforcement chain, comprising: the process control plate and the data storage center respectively output safety regulation and data information as data sources to be sent to the reinforcement learning model of the underlying system; and generating reinforced safety rules and encrypted data information according to the reinforced learning model and the data source, and feeding back the reinforced safety rules and the encrypted data information to the process control plate and the data storage center respectively to form a convolution enhancement chain together with the blockchain network.
In one embodiment of the invention, the generating the enhanced security rules and the encrypted data information according to the enhanced learning model and the data source includes: performing iterative operation by taking the data source as input continuously based on the reinforcement learning model, calculating a Q value which causes state update after each action conversion, and performing accumulation operation on the Q value obtained each time to obtain an accumulated Q value; encrypting the learning data through secure multi-party calculation, taking the encrypted learning data as an action space A of the reinforcement learning model, and participating in action and state conversion; inputting the accumulated Q value into the action space A to perform exploration comparison of executed actions and existence actions, measuring whether the executed actions are needed to be explored or not, and inputting the accumulated Q value into the environment from the action space A to obtain the rewarded value after the action conversion; and iteratively generating an adaptive safety rule according to the reward value, incorporating the adaptive safety rule into a safety rule pool, guiding the state change according to the environment transition, transmitting the state change into a state space S, and feeding back the state change to the initial reinforcement learning model.
According to a second aspect of the present invention, there is also provided a data privacy protection apparatus based on a convolution enhancement chain, comprising: the data node establishing module is used for acquiring the learning data corresponding to each learning subject, establishing corresponding private data nodes according to the learning subjects and storing the learning data into the corresponding private data nodes; the network connection establishment module is used for establishing a secure network connection between the private data nodes based on an intelligent contract and a consensus protocol of the blockchain network and realizing self state update by mutual communication; and the privacy protection module is used for constructing a bottom layer system based on reinforcement learning and security calculation at each private data node, forming a convolution enhancement chain together with the blockchain network, and carrying out privacy protection on the learning data.
In one embodiment of the present invention, the network connection establishment module is specifically configured to: a process control plate and a data storage center are respectively arranged in the private data node to respectively control two processes of a data input node and a data output node; and constructing a blockchain network bridge between the process control plate and the data storage center, wherein the blockchain network bridge is used for receiving response information of other private data nodes in real time and updating self state information according to the response information.
In one embodiment of the present invention, the privacy protection module is specifically configured to: the process control plate and the data storage center respectively output safety regulation and data information as data sources to be sent to the reinforcement learning model of the underlying system; and generating reinforced safety rules and encrypted data information according to the reinforced learning model and the data source, and feeding back the reinforced safety rules and the encrypted data information to the process control plate and the data storage center respectively to form a convolution enhancement chain together with the blockchain network.
In one embodiment of the present invention, the privacy protection module is specifically configured to: performing iterative operation by taking the data source as input continuously based on the reinforcement learning model, calculating a Q value which causes state update after each action conversion, and performing accumulation operation on the Q value obtained each time to obtain an accumulated Q value; encrypting the learning data through secure multi-party calculation, taking the encrypted learning data as an action space A of the reinforcement learning model, and participating in action and state conversion; inputting the accumulated Q value into the action space A to perform exploration comparison of executed actions and existence actions, measuring whether the executed actions are needed to be explored or not, and inputting the accumulated Q value into the environment from the action space A to obtain the rewarded value after the action conversion; and iteratively generating an adaptive safety rule according to the reward value, incorporating the adaptive safety rule into a safety rule pool, guiding the state change according to the environment transition, transmitting the state change into a state space S, and feeding back the state change to the initial reinforcement learning model.
According to a third aspect of the present invention there is also provided an electronic device comprising at least one processing unit and at least one storage unit, wherein the storage unit stores a computer program which, when executed by the processing unit, causes the processing unit to perform the steps of the method according to any of the embodiments described above.
According to a fourth aspect of the present invention there is also provided a computer readable storage medium storing a computer program executable by an access authentication device, the computer program, when run on the access authentication device, causing the access authentication device to perform the steps of the method of any of the embodiments described above.
In general, compared with the prior art, the above technical solutions conceived by the present invention can achieve at least the following beneficial effects:
1) By setting a plurality of independent private data nodes to carry out communication interaction, data exchange and state update in a blockchain network, the data stored in the nodes have the characteristic that the blockchain is not easy to tamper in a distributed storage environment, when one node is unexpected, the state of the other nodes can be synchronously restored to the state before being destroyed, and the data information and the security rules in the nodes are not influenced;
2) The protection of the intelligent and network security is added on the basis of the original blockchain service, the security rules are continuously updated and the data information is encrypted, so that the protection degree of the private data nodes in the whole network architecture on threats such as network attack and the like is greatly improved, and the possibility of theft and embezzlement of the data information is also fundamentally prevented.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are required to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a data privacy protection method based on a convolution enhancement chain according to an embodiment of the present application;
FIG. 2 is a schematic diagram of a block chain network communication architecture according to an embodiment of the present disclosure;
fig. 3 is a schematic diagram of a private data node structure provided in an embodiment of the present application;
FIG. 4 is a schematic diagram of an underlying system architecture according to an embodiment of the present disclosure;
fig. 5 is a schematic diagram of a convolutional enhanced chain network architecture according to an embodiment of the present application;
fig. 6 is a schematic structural diagram of a data privacy protection device based on a convolution enhancement chain according to an embodiment of the present application.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention. In addition, the technical features of the embodiments of the present invention described below may be combined with each other as long as they do not collide with each other.
The terms first, second, third and the like in the description and in the claims of the application and in the above-described figures, are used for distinguishing between different objects and not necessarily for describing a particular sequential or chronological order. Furthermore, the terms "comprise" and "have," as well as any variations thereof, are intended to cover a non-exclusive inclusion. For example, a process, method, system, article, or apparatus that comprises a list of steps or elements is not limited to only those listed steps or elements but may include other steps or elements not listed or inherent to such process, method, article, or apparatus.
As shown in fig. 1, a first embodiment of the present invention proposes a data privacy protection method based on a convolution enhancement chain, for example, including: step S1, learning data corresponding to each learning subject are obtained, corresponding private data nodes are established according to the learning subjects, and the learning data are stored in the corresponding private data nodes; step S2, establishing a secure network connection between the private data nodes based on an intelligent contract and a consensus protocol of a blockchain network, and mutually communicating to realize self-state update; and S3, constructing a bottom layer system based on reinforcement learning and security calculation in each private data node, forming a convolution enhancement chain together with the blockchain network, and carrying out privacy protection on the learning data.
In step S1, each of the learning subjects includes, for example, learning subjects such as teachers, students, parents, schools, etc., and each corresponding private data node is established according to a different learning subject, and relevant data information is stored in the corresponding private data node.
In step S2, in conjunction with the illustration of fig. 2, all private data nodes are linked, for example, based on the blockchain network S10, and a related smart contract S101 and consensus protocol S102 are deployed in the network to cooperatively maintain the secure operation of the entire blockchain network S10 and the status updates of the individual private data nodes S20.
Referring to fig. 3, the private data node S20 in this embodiment is greatly different from the block structure in the standard blockchain, and is not the structure of the block header and the block body any more, but the private data center is constructed to make breakthrough modification on the inside of the node, and the specific structure and deployment are as follows:
and constructing a blockchain network bridge S201 at the head position of the private data node S20, wherein the blockchain network bridge S201 is used for connecting all the private data nodes S20 and performing data information and security rule interaction to realize state transformation.
Meanwhile, a process control block S202 is established at the middle position of the private data node S20 and is used for controlling the data input S207, preventing unknown risk attacks and resisting known network attacks, and recording dangerous access behaviors.
A data storage center S203 is created beside the process control board S202, and is configured to store data information of the private data node S20, and control the data output S208 to other private data nodes to complete the state update, and may also interact with the process control board S207.
In addition, an underlying system S204 is established at the bottom of the private data node S20, for generating safer security rules and performing security processing on data information, and accepting the security rules S205 from the process control block S202 and the data information S206 from the data storage center S203, for enhancing the intellectualization of the system.
Furthermore, the data interaction between private data nodes is separated from the account transaction communication between the existing blockchain networks based on the blockchain networks, so that the conversion of text transmission channels with content properties and state information of the nodes, including states of data information, process control rules and the like, is increased, and the data interaction is ensured to be carried out under a safe and reliable network.
Specifically, the private data nodes communicate with each other, for example, with their own identity attributes. Issuing a private digital signature p for each private data node through the embedding of blockchain network bridge S201 in private data node S20 j The personal portrait and public and private keys of the main body are contained in the personal portrait and the public and private keys for information interaction and identity authentication. Based on digital signature p of each node j To authenticate the transmitted request or the transmitted information, and at the same time, the public keys p of all nodes a Are all publically propagated in the blockchain network S10, and only nodes based on the blockchain network can receive the public keys p of other nodes a And uses it to unlock the user's private key encrypted information, thereby accepting the request and data information.
For example, the implementation steps are as follows:
(1) Assume that data node a sends a network communication request to node B and connects to a blockchain network;
(2) The blockchain network will issue digital certificates for nodes a and B as l a And l b And publishes public keys p of nodes a and B, respectively, in the network a And p b ；
(3) Data node a passes its own private key s on a request sent to node B a Encrypting the content, and simultaneously adding a digital signature which is unique to the node A to the request;
(4) The block chain network performs verification processing on the data signature and the CA certificate of the data request, and if the verification is passed, the data request is allowed to be transmitted; if the check fails, the certificate is proved to be forged, the communication request is refused, and the whole network notifies the node to be a malicious node;
(5) After receiving the communication request from the node A, the data node B randomly generates a symmetric encryption key locally, can check the attribution of a communication party through a digital signature, and can check the communication content through a public key disclosed in a blockchain network;
(6) Finally, the secure transmission between the data and the state can be carried out on the blockchain network through the symmetrical encryption mode among the private data nodes.
Specifically, for example, encryption processing is performed on student growth data information by secure computation, assuming that node a, the given data is x= (x) 1 ,x 2 ,x 3 ,…,x n ) According to private key s of node A by Enc encryption algorithm a The data are processed, and the formula is as follows:
y＝Enc(x,s a )；
wherein y refers to ciphertext obtained by encrypting data, and then passing through a Dec decryption algorithm and p a The ciphertext y is restored according to the following formula:
x＝Dex(y,p a )；
based on the above, the process of encrypting and decrypting the student growth data is completed, and privacy protection processing for the data is realized.
In step S3, for example, the security rule S205 output by the process control board and the data information S206 in the data store are input into the reinforcement learning and security calculation based system S204, and then a more powerful security rule S205 is generated at the system and applied to the process control board S202, and the encrypted data information S206 is fed back to the data storage center S203, thereby constructing a convolution enhancement chain together with the blockchain network S10.
Specifically, the process control plate is a control module built in the private data node, which is different from the control instruction of the existing invariable operating system, and the process control mode is continuously improved and intelligent through the continuously strengthened derived security rule, so that the security inspection of data can be realized, the data node is protected from external attack, the resistance of the node to dangerous attack is improved, and good data privacy protection is realized.
FIG. 4 is a diagram of a network architecture of an underlying system in which reinforcement learning and security computation are combined to form a system architecture that is commonly employed. And carrying out model-free learning on a small amount of marks and a large amount of marked data through reinforcement learning, and continuously iterating out strong security rules to adapt to unknown network risks. Secondly, the security calculation algorithm can encrypt the student data information stored in the data storage center, so that the data is not indicated that the privacy is also revealed after the data is revealed. The specific operation steps are as follows:
first, learning data and initialized safety rules are input into a reinforcement learning model of an underlying system as a data source.
And secondly, performing iterative operation by taking learning data and a safety rule as input continuously based on a reinforcement learning model, calculating a Q value which causes state update after each action conversion, and performing accumulation operation on the Q value based on the strategy until the maximum accumulated Q value is obtained.
And thirdly, encrypting the input data through secure multiparty calculation, and taking the encrypted data as an action space A of the reinforcement learning model to participate in action-state conversion.
And fourthly, inputting an action space according to the accumulated Q value and the optimal strategy obtained in the second step, and comparing the executed action with the existing action to judge whether the executed action is still to be explored. And input the action space into the environment to obtain the rewarding value after the action transformation.
And fifthly, iteratively generating an adaptive safety rule based on the steps, incorporating the adaptive safety rule into a safety rule pool, guiding the state to change according to the environment transition, transferring the state to a state space S, and feeding back to the initial reinforcement learning model.
According to the steps, the operation of the bottom layer system can be completed, and the current most powerful safety rule is generated and put into the process control plate to realize system management and operation.
For example, assume that the data input is x= (x) 1 ,x 2 ,x 3 ,…,x n ) Before data enter the data storage center of each node, the data need to flow through the process control plate; the process control plate passes through the initialized safety rule r 1 Checking, namely capturing the safety and the harmfulness of the data input, judging the data input to be a network attack or illegal input risk if relevant risk loopholes exist, and recording the data input to be A if the data input is a network abnormal attack i ∈A＝{A 1 ,A 2 ,…,A n If it is illegally input, it is recorded as P i ∈P＝{P 1 ，P 2 ,…,P m -a }; finally, the noted network attacks are delivered as input into the underlying system architecture as reinforcement learning rewards mechanisms until security rules are generated that can be completed to combat these risks.
In the reinforcement learning model, initial states are(s) t ，a t ，s t+1 ，r t )，s t Representing the current state of the model, here representing the safety rule S205 in the current process control block S202, and simultaneously being input together as a super parameter; a, a t Representing an action performed in the current state, the security rule S205 is continuously updated by learning the data information S206 from the data storage center S203, each learning being an action performed once; s is(s) t+1 Representing a change in state after an action is performed; r is (r) t Representing the rewards available after the current series of operations, the definition of the rewards function is as follows:
wherein, alpha is a balance factor for controlling the balance state between the acquired network anomaly attack and the detected illegal input;
by means ofThe epsilon-greedy strategy iterates the above procedure until a policy pi-based is obtained * The security rules under(s) can achieve the best effect, we will also base policy pi * (s) becomes the optimal strategy. The E-greedy policy algorithm is as follows:
wherein mu is [0,1 ]]，n s,max Is the state explored in reinforcement learning, n s,t Is the number of states currently explored, ε min ，ε max Is the minimum and maximum values of the exploration percentages of the epsilon-greedy algorithm;
according to the above, the selection of the reinforcement learning epsilon-greedy algorithm and the selection of the reward function are defined as follows for the Q value function:
Q(s t ,a t )＝Q(s t ,a t )+α(r+γQ(s t+1 ,a t+1 ；θ)-Q(s t ,a t )) (3)
wherein, gamma is E [0,1 ]]Is a discount factor for controlling the influence of future rewards in the current state, θ represents the weight parameter of the convolutional neural network, and r belongs to the current state and performs a t Rewards after action;
meanwhile, the loss function training for the Q function is as follows:
wherein y is i Is the object of the method consisting of Q (s t ,a t ) Is calculated from the reward and the next state s t+1 Estimating the ground;
according to the conditions, the optimal strategy selection is carried out, and the definition is as follows:
π * (s)＝argmaxQ * (s,a) (5)
wherein Q is * (s, a) represents an optimal Q-value function.
The data storage center is built in the data node and is mainly responsible for the data storage and extraction functions, risk management and control can be continuously iterated on the data in the data warehouse through interaction with the process control plate, and each time the safety rule is updated, all the data can be checked again in the private data node, the data are continuously prevented and controlled, the safety of the node and the blockchain network is improved, and therefore learning subject data are better protected.
The intelligent underlying system is built based on reinforcement learning and security calculation, and the reinforcement learning mechanism is utilized to continuously improve the initially built security rules so as to resist the risk. The safety rule is selected mainly by judging the Q value, and the judgment formula is as follows:
wherein, the liquid crystal display device comprises a liquid crystal display device,refers to the rewards of executing the a action in the s state, gamma is a discount factor for judging the future conversion of rewards in the current state,/>Refers to the state transition probability, the probability of performing an a-state transition to an s' -state in the current s-state, v π (s ') refers to the accumulation of all Q values for an action performed in accordance with the pi policy mechanism in the s' state, which also includes training the θ parameter.
Meanwhile, for training of the convolution enhancement chain, the convergence theta is removed by a gradient descent method mainly based on the weight parameter theta of the convolution neural network, and meanwhile, the target value y can be reduced i And Q value Q(s) t At), the gradient of the loss function with respect to the neural network is as follows:
wherein, the liquid crystal display device comprises a liquid crystal display device,gradient representing loss function ∈ ->Expressed in a specific weight parameter theta i Gradient of Q value.
Based on the above, the encryption processing of the data information input to the underlying system by the data storage center through the secure computation comprises the following steps:
the security computation may be abstracted into a mathematical model with the following formula:
f(x 1 ,x 2 ,…,x n )＝(y 1 ,y 2 ,…,y n ) (6)
data information of the data storage center is set to x= { X 1 ,x 2 ,…,x n Any one data information x i Through f i (x)＝a i +b i x+c i x 2 An encryption process is performed while the public key (p a ,p b ) Performing encryption operation;
wherein f i (x) Splitting may be performed at appropriate times to split the encrypted information into multiple parts such as f 1 (x),f 2 (x),…,f n (x) Finally, integrating in the data storage center (S203), and superposing other parts, as follows:
F(x)＝f 1 (x)+f 2 (x)+…+f n (x) (7)
the encrypted data information can be restored in the data storage center, and the formula is as follows:
y i ＝f i (x) (8)
the bottom layer system feeds the encrypted data information back to the data storage plate, and continuously propagates and updates the encrypted data in the blockchain network.
In addition, the network architecture of the convolutional enhanced chain formed by the convolutional neural network, the reinforcement learning and the blockchain network is shown in fig. 5, the data source is preprocessed by establishing the convolutional neural network architecture, and the processed data is transmitted to the reinforcement learning model, so that the training efficiency is greatly improved, and the overfitting phenomenon caused by too much learning main body data information is reduced. And finally, the generated reinforced security rules are used as output and are transmitted to a blockchain network, so that the security of the whole network architecture is enhanced, and the whole process of the convolution reinforced chain is formed.
The specific structure and arrangement are as follows:
the first part is used for constructing each layer of the convolutional neural network, the parameter weight theta of the convolutional neural network can be continuously trained along with the gradient drop by continuously flowing through the layers, so that precisely filtered data can be obtained for training of reinforcement learning, and the first part is used for describing the workflow of the convolutional neural network and the flowing sequence among all layers;
a second part for specifically introducing each structure of the convolutional neural network of the first part, and finally obtaining a data stream filtered by the convolutional neural network by inputting data and then passing through an input layer and a hidden layer to an output layer;
part III, introducing reinforcement learning to learn data pre-trained by using a convolutional neural network, and then generating a strong safety rule, wherein the specific flow is shown in figure 4;
according to the above, the generated reinforced safety rule is input into the blockchain network, so that the convolutional neural network, the reinforced learning and the blockchain network are integrated into a whole, and a convolutional reinforced chain with stronger function is formed.
In summary, according to the data privacy protection method based on the convolution enhancement chain provided by the first embodiment of the present invention, by setting a plurality of independent private data nodes to perform communication interaction, data exchange and state update in the blockchain network, the data stored in the nodes has the characteristic that the blockchain is not easy to tamper in the distributed storage environment, when an accident occurs in one of the nodes, the state of the other nodes can be synchronously restored to the state before being destroyed, and the data information and the security rules in the nodes are not affected; the protection of the intelligent and network security is added on the basis of the original blockchain service, the security rules are continuously updated and the data information is encrypted, so that the protection degree of the private data nodes in the whole network architecture on threats such as network attack and the like is greatly improved, and the possibility of theft and embezzlement of the data information is also fundamentally prevented.
In addition, as shown in fig. 6, a second embodiment of the present invention proposes a data privacy protection apparatus 30 based on a convolution enhancement chain, for example, including: a data node establishment module 301, a network connection establishment module 302 and a privacy protection module 303.
The data node establishing module 301 is configured to obtain learning data corresponding to each learning subject, establish corresponding private data nodes according to the learning subjects, and store the learning data into the corresponding private data nodes. The network connection establishment module 302 is configured to establish a secure network connection between the private data nodes based on the intelligent contracts and the consensus protocol of the blockchain network, and communicate with each other to implement self-state update. The privacy protection module 303 is configured to construct an underlying system based on reinforcement learning and security computation at each private data node, and form a convolution enhancement chain together with the blockchain network, so as to perform privacy protection on the learning data.
It should be noted that the method implemented by the data privacy protection apparatus 30 based on the convolution enhancement chain according to the second embodiment of the present invention is as described in the foregoing first embodiment, so that detailed description thereof will not be provided herein. Optionally, each module and the other operations or functions in the second embodiment are respectively to implement the data privacy protection method based on the convolutional enhanced chain in the first embodiment, and the beneficial effects of this embodiment are the same as those of the foregoing first embodiment, which is not repeated herein for brevity.
The third embodiment of the present invention also proposes an electronic device, for example, including: at least one processing unit, and at least one storage unit, wherein the storage unit stores a computer program, which when executed by the processing unit, causes the processing unit to perform the method according to the first embodiment, and the beneficial effects of the electronic device provided by the present embodiment are the same as those of the data privacy protection method based on the convolution enhanced chain provided by the first embodiment.
The third embodiment of the present invention also provides a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the above method, and the beneficial effects of the computer readable storage medium provided by the present embodiment are the same as those of the data privacy protection method based on a convolution enhancement chain provided by the first embodiment.
The computer readable storage medium may include, among other things, any type of disk including floppy disks, optical disks, DVDs, CD-ROMs, micro-drives, and magneto-optical disks, ROM, RAM, EPROM, EEPROM, DRAM, VRAM, flash memory devices, magnetic or optical cards, nanosystems (including molecular memory ICs), or any type of media or device suitable for storing instructions and/or data.
It should be noted that, for simplicity of description, the foregoing method embodiments are all expressed as a series of action combinations, but it should be understood by those skilled in the art that the present application is not limited by the order of actions described, as some steps may be performed in other order or simultaneously in accordance with the present application. Further, those skilled in the art will also appreciate that the embodiments described in the specification are all preferred embodiments, and that the acts and modules referred to are not necessarily required in the present application.
In the foregoing embodiments, the descriptions of the embodiments are emphasized, and for parts of one embodiment that are not described in detail, reference may be made to related descriptions of other embodiments.
In the several embodiments provided in this application, it should be understood that the disclosed apparatus may be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative, such as the division of the units, merely a logical function division, and there may be additional manners of dividing the actual implementation, such as multiple units or components may be combined or integrated into another system, or some features may be omitted, or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be through some service interface, device or unit indirect coupling or communication connection, electrical or otherwise.
The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional unit in each embodiment of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.
The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable memory. Based on such understanding, the technical solution of the present application may be embodied in essence or a part contributing to the prior art or all or part of the technical solution in the form of a software product stored in a memory, including several instructions for causing a computer device (which may be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in the embodiments of the present application. And the aforementioned memory includes: a U-disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), a removable hard disk, a magnetic disk, or an optical disk, or other various media capable of storing program codes.
Those of ordinary skill in the art will appreciate that all or a portion of the steps in the various methods of the above embodiments may be performed by hardware associated with a program that is stored in a computer readable memory, which may include: flash disk, read-Only Memory (ROM), random-access Memory (Random Access Memory, RAM), magnetic or optical disk, and the like.
The foregoing is merely exemplary embodiments of the present disclosure and is not intended to limit the scope of the present disclosure. That is, equivalent changes and modifications are contemplated by the teachings of this disclosure, which fall within the scope of the present disclosure. Embodiments of the present disclosure will be readily apparent to those skilled in the art from consideration of the specification and practice of the disclosure herein. This application is intended to cover any adaptations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a scope and spirit of the disclosure being indicated by the claims.
The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
It will be readily appreciated by those skilled in the art that the foregoing description is merely a preferred embodiment of the invention and is not intended to limit the invention, but any modifications, equivalents, improvements or alternatives falling within the spirit and principles of the invention are intended to be included within the scope of the invention.
Claims (6)
1. A data privacy protection method based on a convolution enhancement chain, comprising:
acquiring learning data corresponding to each learning subject, establishing corresponding private data nodes according to the learning subjects, and storing the learning data into the corresponding private data nodes;
establishing a secure network connection between the private data nodes based on a block chain network protocol, and mutually communicating to realize self state update;
constructing a bottom layer system based on reinforcement learning and security calculation at each private data node, forming a convolution enhancement chain together with the blockchain network, and performing privacy protection on the learning data;
constructing an underlying system based on reinforcement learning and security computation at each private data node, forming a convolution enhancement chain together with the blockchain network, comprising:
respectively outputting safety regulation and data information by a process control plate and a data storage center as a data source and sending the data information to a reinforcement learning model of the underlying system;
generating reinforced safety rules and encrypted data information according to the reinforced learning model and the data source, and feeding back the reinforced safety rules and the encrypted data information to the process control plate and the data storage center respectively to form a convolution enhancement chain together with the blockchain network;
the generating the reinforced security rules and the encrypted data information according to the reinforced learning model and the data source comprises the following steps:
performing iterative operations with the data source continuously as input based on the reinforcement learning model, and calculating a state update caused after each motion transitionQValue and the value obtained each timeQThe value is accumulated to obtain the accumulationQA value;
encrypting the learning data through secure multi-party calculation, and taking the encrypted learning data as an action space of the reinforcement learning modelAParticipate in actions and state transitions;
integrating the saidQValue input into the motion spaceAComparing the performed action with the search of the existence action, measuring whether the search is still performed or not, and obtaining the action spaceAInputting the result into the environment to obtain the rewarding value after the action transformation;
iteratively generating an adaptive security rule according to the reward value, incorporating the adaptive security rule into a security rule pool, guiding the state change according to the transition of the environment, and introducing the state change into a state spaceSAnd feeding back to the initial reinforcement learning model.
2. The method for protecting data privacy based on convolution enhancement chains according to claim 1, wherein said establishing a secure network connection between each of said private data nodes and communicating with each other to realize self-status update comprises:
a process control plate and a data storage center are respectively arranged in the private data node to respectively control two processes of a data input node and a data output node;
and constructing a blockchain network bridge between the process control plate and the data storage center, wherein the blockchain network bridge is used for receiving response information of other private data nodes in real time and updating self state information according to the response information.
3. A data privacy protection apparatus based on a convolution enhancement chain, comprising:
the data node establishing module is used for acquiring the learning data corresponding to each learning subject, establishing corresponding private data nodes according to the learning subjects and storing the learning data into the corresponding private data nodes;
the network connection establishment module is used for establishing a secure network connection between the private data nodes based on an intelligent contract and a consensus protocol of the blockchain network and realizing self state update by mutual communication;
the privacy protection module is used for constructing a bottom layer system based on reinforcement learning and security calculation at each private data node, forming a convolution enhancement chain together with the blockchain network, and carrying out privacy protection on the learning data;
the privacy protection module is specifically configured to: respectively outputting safety regulation and data information by a process control plate and a data storage center as a data source and sending the data information to a reinforcement learning model of the underlying system; generating reinforced safety rules and encrypted data information according to the reinforced learning model and the data source, and feeding back the reinforced safety rules and the encrypted data information to the process control plate and the data storage center respectively to form a convolution enhancement chain together with the blockchain network;
the privacy protection module is specifically configured to: based on the followingThe reinforcement learning model takes the data source as input to execute iterative operation continuously, and calculates the state update caused by each action transitionQValue and the value obtained each timeQThe value is accumulated to obtain the accumulationQA value; encrypting the learning data through secure multi-party calculation, and taking the encrypted learning data as an action space of the reinforcement learning modelAParticipate in actions and state transitions; integrating the saidQValue input into the motion spaceAComparing the performed action with the search of the existence action, measuring whether the search is still performed or not, and obtaining the action spaceAInputting the result into the environment to obtain the rewarding value after the action transformation; iteratively generating an adaptive security rule according to the reward value, incorporating the adaptive security rule into a security rule pool, guiding the state change according to the transition of the environment, and introducing the state change into a state spaceSAnd feeding back to the initial reinforcement learning model.
4. The data privacy protection apparatus based on convolutional enhancement chain as claimed in claim 3, wherein the network connection establishment module is specifically configured to:
a process control plate and a data storage center are respectively arranged in the private data node to respectively control two processes of a data input node and a data output node;
and constructing a blockchain network bridge between the process control plate and the data storage center, wherein the blockchain network bridge is used for receiving response information of other private data nodes in real time and updating self state information according to the response information.
5. An electronic device comprising at least one processing unit, and at least one storage unit, wherein the storage unit stores a computer program that, when executed by the processing unit, causes the processing unit to perform the steps of the method of any of claims 1-2.
6. A computer readable storage medium, characterized in that it stores a computer program executable by an access authentication device, which when run on the access authentication device causes the access authentication device to perform the steps of the method according to any of claims 1-2.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210646887.6A CN115118462B (en) | 2022-06-09 | 2022-06-09 | Data privacy protection method based on convolution enhancement chain |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210646887.6A CN115118462B (en) | 2022-06-09 | 2022-06-09 | Data privacy protection method based on convolution enhancement chain |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115118462A CN115118462A (en) | 2022-09-27 |
| CN115118462B true CN115118462B (en) | 2023-07-18 |
Family
ID=83325948
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210646887.6A Active CN115118462B (en) | 2022-06-09 | 2022-06-09 | Data privacy protection method based on convolution enhancement chain |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115118462B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116310238B (en) * | 2023-03-16 | 2024-03-22 | 华中师范大学 | Multi-user virtual avatar interaction behavior safety protection method and system |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112866295A (en) * | 2021-03-23 | 2021-05-28 | 黄超 | Block chain big data crawler-prevention processing method and cloud platform system |
| CN113191530A (en) * | 2021-04-09 | 2021-07-30 | 汕头大学 | Block link point reliability prediction method and system with privacy protection function |
| CN113486935A (en) * | 2021-06-24 | 2021-10-08 | 南京烽火星空通信发展有限公司 | Block chain application flow identification method based on DPI and CNN |
| CN114118438A (en) * | 2021-10-18 | 2022-03-01 | 华北电力大学 | Privacy protection machine learning training and reasoning method and system based on block chain |
| CN114189892A (en) * | 2021-12-15 | 2022-03-15 | 北京工业大学 | Cloud-edge collaborative Internet of things system resource allocation method based on block chain and collective reinforcement learning |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109685501B (en) * | 2018-12-04 | 2023-04-07 | 暨南大学 | Auditable privacy protection deep learning platform construction method based on block chain excitation mechanism |
-
2022
- 2022-06-09 CN CN202210646887.6A patent/CN115118462B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112866295A (en) * | 2021-03-23 | 2021-05-28 | 黄超 | Block chain big data crawler-prevention processing method and cloud platform system |
| CN113191530A (en) * | 2021-04-09 | 2021-07-30 | 汕头大学 | Block link point reliability prediction method and system with privacy protection function |
| CN113486935A (en) * | 2021-06-24 | 2021-10-08 | 南京烽火星空通信发展有限公司 | Block chain application flow identification method based on DPI and CNN |
| CN114118438A (en) * | 2021-10-18 | 2022-03-01 | 华北电力大学 | Privacy protection machine learning training and reasoning method and system based on block chain |
| CN114189892A (en) * | 2021-12-15 | 2022-03-15 | 北京工业大学 | Cloud-edge collaborative Internet of things system resource allocation method based on block chain and collective reinforcement learning |
Non-Patent Citations (4)
| Title |
|---|
| 《A Privacy-Preserving-Based Secure Framework Using Blockchain-Enabled Deep-Learning in Cooperative Intelligent Transport System》;Randhir Kumar etal;《IEEE Transactions on Intelligent Transportation Systems》;全文 * |
| 《Blockchain-Enabled Resource Trading and Deep Reinforcement Learning-Based Autonomous RAN Slicing in 5G》;Gordon Owusu Boateng etal;IEEE;全文 * |
| 如何保护数据安全和个人隐私;李子臣;杨义先;;金融经济(11);全文 * |
| 面向深度学习的差分隐私保护方法;芈小龙;隋景鹏;;舰船电子工程(09);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115118462A (en) | 2022-09-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Cao et al. | Cloud-assisted secure eHealth systems for tamper-proofing EHR via blockchain | |
| Fan et al. | Dredas: Decentralized, reliable and efficient remote outsourced data auditing scheme with blockchain smart contract for industrial IoT | |
| Sarmah | Understanding blockchain technology | |
| Lee et al. | Digestive neural networks: A novel defense strategy against inference attacks in federated learning | |
| CN107147720B (en) | Traceable effective public auditing method and traceable effective public auditing system in cloud storage data sharing | |
| CN112232527A (en) | Safe distributed federal deep learning method | |
| Tran et al. | A survey on privacy-preserving blockchain systems (PPBS) and a novel PPBS-based framework for smart agriculture | |
| Cheng et al. | Polynomial-based modifiable blockchain structure for removing fraud transactions | |
| Muheidat et al. | Artificial intelligence and blockchain for cybersecurity applications | |
| Zhang et al. | Security and privacy threats to federated learning: Issues, methods, and challenges | |
| Lyu et al. | Towards fair and decentralized privacy-preserving deep learning with blockchain | |
| Yang et al. | Security of federated learning for cloud‐edge intelligence collaborative computing | |
| CN115118462B (en) | Data privacy protection method based on convolution enhancement chain | |
| Wang et al. | Insider collusion attack on privacy-preserving kernel-based data mining systems | |
| Garman et al. | Rational zero: Economic security for zerocoin with everlasting anonymity | |
| CN115687758A (en) | User classification model training method and user detection method | |
| Shah et al. | On the vital aspects and characteristics of cryptocurrency—A survey | |
| Elisa et al. | A secure and privacy-preserving e-government framework using blockchain and artificial immunity | |
| Khodaiemehr et al. | Navigating the quantum computing threat landscape for blockchains: A comprehensive survey | |
| Prateek et al. | Post‐quantum blockchain–enabled services in scalable smart cities | |
| Alexander et al. | Cybersecurity, information assurance, and big data based on blockchain | |
| Upadhyay et al. | Auditing metaverse requires multimodal deep learning | |
| Gojka et al. | Security in distributed ledger technology: An analysis of vulnerabilities and attack vectors | |
| Xie et al. | A survey on vulnerability of federated learning: A learning algorithm perspective | |
| Maram | Bitcoin generation using Blockchain technology |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |