CN112738061A - Information processing method, device, management platform, electronic equipment and storage medium - Google Patents
Information processing method, device, management platform, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN112738061A CN112738061A CN202011556622.4A CN202011556622A CN112738061A CN 112738061 A CN112738061 A CN 112738061A CN 202011556622 A CN202011556622 A CN 202011556622A CN 112738061 A CN112738061 A CN 112738061A
- Authority
- CN
- China
- Prior art keywords
- information
- node
- storage
- management platform
- storage node
- 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.)
- Granted
Links
Images
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
- H04L63/0435—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 wherein the sending and receiving network entities apply symmetric encryption, i.e. same key used for encryption and decryption
-
- 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
- H04L63/12—Applying verification of the received information
- H04L63/123—Applying verification of the received information received data contents, e.g. message integrity
-
- 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
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1097—Protocols in which an application is distributed across nodes in the network for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L9/00—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols
- H04L9/32—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols including means for verifying the identity or authority of a user of the system or for message authentication, e.g. authorization, entity authentication, data integrity or data verification, non-repudiation, key authentication or verification of credentials
- H04L9/3236—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols including means for verifying the identity or authority of a user of the system or for message authentication, e.g. authorization, entity authentication, data integrity or data verification, non-repudiation, key authentication or verification of credentials using cryptographic hash functions
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L9/00—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols
- H04L9/50—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols using hash chains, e.g. blockchains or hash trees
Abstract
The application provides an information processing method, an information processing device, a management platform, an electronic device and a storage medium, wherein the method comprises the following steps: sending a parameter information acquisition instruction of the target model to a management platform; receiving information of at least two storage nodes returned by the management platform based on the instruction; the at least two storage nodes respectively store parameter information of different network model layers in the target model; the at least two storage nodes belong to the management platform; and acquiring the parameter information of the network model layer stored at each storage node according to the information of each storage node so as to realize model sharing under the condition of ensuring the safety of the model.
Description
Technical Field
The present application relates to the field of information security technologies, and in particular, to an information processing method, an information processing apparatus, a management platform, an electronic device, and a storage medium.
Background
With the development of internet technology, which is an era centering on data in the future, the industries are gradually fused, and Artificial Intelligence (AI) models are more and more widely used, however, in order to obtain a trained AI model, a large amount of manpower and material resources are needed to collect, clean and train data repeatedly to obtain parameter information of each network model layer in the AI model, so that the sharing of the model is particularly important, in the prior art, the complete relevant information of the model is usually recorded in a file and uploaded to a server for storage, so that the model can be obtained by the model demander directly downloading the file via the server, however, the model storage mode can cause the model to be easily obtained illegally, and the safety is poor, so that the problem of realizing model sharing under the condition of ensuring the safety of the model is to be solved.
Disclosure of Invention
In view of this, an object of the embodiments of the present application is to provide an information processing method, an information processing apparatus, a management platform, an electronic device, and a storage medium, so as to implement model sharing under the condition that model security is ensured.
In a first aspect, an embodiment of the present application provides an information processing method, where the method includes: sending a parameter information acquisition instruction of the target model to a management platform; receiving information of at least two storage nodes returned by the management platform based on the instruction; the at least two storage nodes respectively store parameter information of different network model layers in the target model; the at least two storage nodes belong to the management platform; and acquiring the parameter information of the network model layer stored at each storage node according to the information of each storage node.
In the implementation process, the parameter information acquisition instruction of the target model is sent to the management platform, and according to the information of at least two storage nodes returned by the management platform, the parameter information related to the target model is accurately acquired from the storage node, so that the sharing of the model is realized.
In a possible design based on the first aspect, the receiving information of at least two storage nodes returned by the management platform based on the instruction includes: receiving the notary node information returned by the management platform based on the instruction; the notary nodes belong to the management platform, and each notary node in the management platform corresponds to a model one to one; sending a node information request to the notary node according to the information of the notary node; and receiving the information of the at least two storage nodes returned by the notary node based on the node information request.
In the implementation process, each notary node in the management platform corresponds to the model one to one, and the notary node manages the information of the storage node corresponding to the model corresponding to the notary node, so that the load pressure of the management platform is reduced.
In a possible design based on the first aspect, the information of the at least two storage nodes includes: a public key corresponding to each storage node in the at least two storage nodes; the acquiring parameter information of the network model layer stored at each storage node according to the information of each storage node includes: sending a parameter request to each storage node according to the information of the storage node; wherein the parameter request includes: a public key corresponding to the storage node; receiving a feedback result returned by the storage node based on the parameter request; the feedback result comprises: an encrypted key, and encrypted parameter information; decrypting the encrypted key by using the corresponding public key to obtain a symmetric key; and decrypting the encrypted parameter information by using the symmetric key to obtain the parameter information of the corresponding network model layer.
In the implementation process, the parameter information of each network model layer of the target model is encrypted, the key used for encrypting the parameter information of each network model layer is encrypted and then stored in the corresponding storage node, so that the safety of the parameter information of the target model can be further ensured, the encrypted key is decrypted by using the public key corresponding to each storage node to obtain a symmetric key, the encrypted parameter information is decrypted by using the symmetric key, the parameter information of the corresponding network model layer can be obtained, and the model sharing is realized.
In a possible design based on the first aspect, the information of the at least two storage nodes further includes: an identifier of a network model layer corresponding to each of the at least two storage nodes; the method further comprises the following steps: and according to the identifier of the network model layer corresponding to each storage node, utilizing the parameter information of the network model layer acquired from each storage node to correspondingly set the value of each parameter in the initial model of the target model, so as to obtain the target model.
In the implementation process, according to the identifier of the network model layer corresponding to each storage node, the parameter information of the network model layer acquired from each storage node is used to correspondingly set the value of each parameter in the initial model of the target model, so as to accurately obtain the target model.
In a second aspect, an embodiment of the present application provides an information processing method, which is applied to a management platform, and the method includes: receiving a parameter information acquisition instruction of a target model; determining information of at least two storage nodes according to the instruction; the at least two storage nodes respectively store parameter information of different network model layers in the target model; the at least two storage nodes belong to the management platform; and sending the information of the at least two storage nodes to the instruction sender.
In the implementation process, after receiving a parameter information acquisition command of the target model, the information of at least two storage nodes is returned to a command sender, the at least two storage nodes respectively store parameter information of different network model layers in the target model, so as to ensure that the instruction sender can obtain all parameter information of the target model from at least two storage nodes, realize the sharing of the model, this way by storing parameter information of different network model layers in the model with different storage nodes, even if an illegal acquirer acquires part of the parameter information of the model from one storage node, it is not easy to acquire all the parameter information of the model, therefore, compared with the prior art that all information of the model is stored by directly utilizing one storage node, the model can be better prevented from being illegally acquired, and the safety of the model is improved.
Based on the second aspect, in one possible design, the management platform further includes: a notary node; the method further comprises the following steps: aiming at each storage node in the at least two storage nodes, the notary node continuously receives the Hash value sent by the storage node; the Hash value is determined by the storage node by using the parameter information of the network model layer stored in the storage node at the current moment; the notary node determines that the parameter information stored on the storage node changes when the Hash value received at the current moment is different from the Hash value stored in advance; the pre-stored Hash value is determined by the storage node by using the parameter information of the network model layer of the target model sent by the management platform; the notary node generates prompt information; and the prompt information is used for prompting that the parameter information stored on the storage node is tampered.
In the implementation process, each storage node determines a Hash value by using the parameter information of the network model layer stored in the storage node, and when the stored parameter information changes, the determined Hash value also changes, so that for each storage node in the at least two storage nodes, a notary node generates prompt information when determining that the Hash value sent by the storage node and received at the current moment is different from the Hash value stored in advance, so that when the parameter information stored in the storage node is tampered, a manager can be prompted in time, and the manager can process the parameter information stored in the storage node in time.
In a third aspect, an embodiment of the present application provides an information processing apparatus, including: the instruction sending unit is used for sending a parameter information acquisition instruction of the target model to the management platform; the receiving unit is used for receiving the information of the at least two storage nodes returned by the management platform based on the instruction; the at least two storage nodes respectively store parameter information of different network model layers in the target model; the at least two storage nodes belong to the management platform; and the acquisition unit is used for acquiring the parameter information of the network model layer stored at each storage node according to the information of each storage node.
Based on the third aspect, in a possible design, the receiving unit is specifically configured to receive the information of the notary node returned by the management platform based on the instruction; wherein the notary node belongs to the management platform; sending a node information request to the notary node according to the information of the notary node; and receiving the information of the at least two storage nodes returned by the notary node based on the node information request.
In a possible design based on the third aspect, the information of the at least two storage nodes includes: a public key corresponding to each storage node in the at least two storage nodes; the acquiring unit is specifically configured to send a parameter request to each storage node according to information of the storage node; wherein the parameter request includes: a public key corresponding to the storage node; receiving a feedback result returned by the storage node based on the parameter request; the feedback result comprises: an encrypted key, and encrypted parameter information; decrypting the encrypted key by using the corresponding public key to obtain a symmetric key; and decrypting the encrypted parameter information by using the symmetric key to obtain the parameter information of the corresponding network model layer.
In a possible design based on the third aspect, the information of the at least two storage nodes further includes: an identifier of a network model layer corresponding to each of the at least two storage nodes; the device further comprises: and according to the identifier of the network model layer corresponding to each storage node, utilizing the parameter information of the network model layer acquired from each storage node to correspondingly set the value of each parameter in the initial model of the target model, so as to obtain the target model.
In a fourth aspect, an embodiment of the present application provides a management platform, where the management platform is configured to receive a parameter information obtaining instruction of a target model; determining information of at least two storage nodes according to the instruction; the at least two storage nodes respectively store parameter information of different network model layers in the target model; the at least two storage nodes belong to the management platform; and sending the information of the at least two storage nodes to the instruction sender.
In a possible design based on the fourth aspect, the management platform further includes: a notary node; the notary node is used for continuously receiving the Hash value sent by each storage node in the at least two storage nodes; the Hash value is determined by the storage node by using the parameter information of the network model layer stored in the storage node at the current moment; when the Hash value received at the current moment is determined to be different from the Hash value stored in advance, determining that the parameter information stored on the storage node changes; the pre-stored Hash value is determined by the storage node by using the parameter information of the network model layer of the target model sent by the management platform; generating prompt information; and the prompt information is used for prompting that the parameter information stored on the storage node is tampered.
In a fifth aspect, an embodiment of the present application provides an electronic device, including a processor and a memory connected to the processor, where a computer program is stored in the memory, and when the computer program is executed by the processor, the electronic device is caused to perform the method of the first aspect or the second aspect.
In a sixth aspect, embodiments of the present application provide a storage medium, in which a computer program is stored, and when the computer program runs on a computer, the computer is caused to execute the method of the first aspect or the second aspect.
Additional features and advantages of the present application will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by the practice of the embodiments of the present application. The objectives and other advantages of the application may be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
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, it should be understood that the following drawings only illustrate some embodiments of the present application and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained from the drawings without inventive effort.
Fig. 1 is a schematic connection diagram of a storage node and a notary node provided in an embodiment of the present application.
Fig. 2 is a schematic flowchart of an information processing method according to an embodiment of the present application.
Fig. 3 is a schematic structural diagram of an information processing apparatus according to an embodiment of the present application.
Fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
Icon: 310-an instruction issue unit; 320-a receiving unit; 330-an obtaining unit; 400-an electronic device; 401-a processor; 402-a memory; 403-communication interface.
Detailed Description
The technical solution in the embodiments of the present application will be described below with reference to the drawings in the embodiments of the present application.
It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, it need not be further defined and explained in subsequent figures. Meanwhile, in the description of the present application, the terms "first", "second", and the like are used only for distinguishing the description, and are not to be construed as indicating or implying relative importance.
For ease of understanding, the manner in which parameter information of the model needs to be stored is described below.
After a management platform receives a target model, determining parameter information of each network model layer in the target model, and sending the parameter information of the network model layer to a storage node which does not store the parameter information of the model in the management platform aiming at each network model layer, wherein the storage node stores the received parameter information; the management platform establishes and stores a corresponding relation between the target model and information of a storage node used for storing parameter information of the target model.
As an embodiment, for each network model layer, the parameter information of the network model layer and the identifier of the network model layer are sent to a storage node in the management platform, where the parameter information of the model is not stored, and the storage node stores the received parameter information and the identifier of the network model layer.
In this embodiment, the information of the storage node includes: an Internet Protocol (IP) address of an interconnection between networks of the storage nodes; in other embodiments, the information of the storage node may further include: and the identifier of the network model layer corresponding to the storage node. In this embodiment, the identifier of the network model layer may use 1,2,3 or other characters to indicate the position ordering of the network model layer in the target model.
In one embodiment, after the storage node receives the parameter information of the network model layer, the storage node stores the parameter information of the network model layer based on an Inter Planet File System (IPFS) protocol.
As an embodiment, for each storage node, after the storage node receives parameter information of a network model layer, the storage node processes the parameter information of the network model layer according to an asymmetric key algorithm to obtain an asymmetric key pair, where the asymmetric key pair includes: a public key and a private key; the storage node encrypts the parameter information of the network model layer by using a randomly generated symmetric key to obtain encrypted parameter information, and encrypts the symmetric key by using a private key in the asymmetric key pair to obtain an encrypted key; the encrypted key, the encrypted parameter information and the asymmetric key pair are then stored using an IPFS-based protocol, it being understood that the public key of the asymmetric key pair stored in the storage node is the information of the storage node.
In one embodiment, the management platform sends all information of the storage nodes used for storing the parameter information of the target model to a predetermined notary node, and the notary node stores the received information of the storage nodes. Wherein the notary node belongs to the management platform. The notary nodes correspond to the models one to one, namely one notary node is used for storing information of each storage node corresponding to one model.
As an implementation manner, the management platform sends the information of the notary node to each storage node for storing parameter information of a target model, and after receiving the information of the notary node, each storage node sends the information of the storage node to the notary node according to the information of the notary node. The information of the storage node may be an IP address of the storage node; the information of the storage node can also be a public key; the information of the storage node may further include: the storage node stores an identification of the network model layer.
As an implementation, the storage node and notary node referred to in the above embodiments may be nodes in a block chain. The notary node stores a first asymmetric key pair and a symmetric key randomly generated by the notary node into a block of the notary node, and sends a public key in the first asymmetric key pair to the at least two storage nodes for storage.
Each storage node determines a Hash value in a block head of the storage node according to the stored parameter information of the network model layer, and the management platform connects the at least two storage nodes with the notary node in series in a Hash addressing mode according to the network model layer corresponding to each storage node for storing the target model and the notary node corresponding to the target model.
As shown in fig. 1, assuming that the number of the at least two storage nodes is two, the specific manner of connecting the at least two storage nodes and the notary node in series is as follows: setting the Hash value in the block tail of the storage node corresponding to the second network model layer of the target model to be the same as the Hash value (Hash1) in the block head of the storage node corresponding to the first network model layer, and pointing the pointer of the storage node corresponding to the second network model layer to the block head of the storage node corresponding to the first network model layer; and setting the Hash value in the block tail of the notary node to be the same as the Hash value (Hash2) in the block head of the storage node corresponding to the second network model layer, and pointing the pointer of the notary node to the block head of the storage node corresponding to the second network model layer, wherein the Hash value of the block head of the notary node is Hash M.
Referring to fig. 2, fig. 2 is a flowchart of an information processing method according to an embodiment of the present application, and the flowchart shown in fig. 2 will be described in detail below, where the method includes the steps of: s21, S22, S23, S24, S25 and S26.
S21: and sending a parameter information acquisition instruction of the target model to the management platform.
S22: the instructions are received.
S23: determining information of at least two storage nodes according to the instruction; the at least two storage nodes respectively store parameter information of different network model layers in the target model; the at least two storage nodes belong to the management platform.
S24: and sending the information of the at least two storage nodes to the instruction sender.
S25: and receiving information of the at least two storage nodes.
S26: and acquiring the parameter information of the network model layer stored at each storage node according to the information of each storage node.
The above method is described in detail below.
S21: and the instruction sending direction sends the parameter information acquisition instruction of the target model to the management platform.
In an actual implementation process, S21 may be implemented in such a manner that the instruction sender sends a parameter information obtaining instruction of the target model to the management platform based on the IPFS protocol.
Wherein, the instruction may include: identity information of the target model. Wherein the identity information of different models is different.
The instructions may further include: identification information of the instruction sender. Wherein, the identification information corresponding to different instruction senders is different.
S22: the management platform receives the instruction.
In the actual implementation process, the management platform receives the instruction in real time or untimely time.
S23: the management platform determines the information of at least two storage nodes according to the instruction; the at least two storage nodes respectively store parameter information of different network model layers in the target model; the at least two storage nodes belong to the management platform.
If the management platform only stores the parameter information of one model at the current time, S23 may be implemented as follows, and the management platform acquires information of at least two storage nodes stored in advance and storing the parameter information of the model according to the instruction.
If the management platform only stores the parameter information of at least two models at the current moment, the instruction comprises: s23 may be implemented in such a manner that the management platform extracts the identity information of the target model from the instruction, and then, according to the identity information of the target model, finds information of a storage node corresponding to the identity information of the target model from a correspondence between the identity information of a prestored model and information of storage nodes, where it is understood that the information of the corresponding storage node is information of the at least two storage nodes.
In this embodiment, the information of the storage node may be an IP address of the storage node.
If the instruction further comprises: as an implementation manner, the S23 may be implemented in such a manner that the management platform extracts the identifier of the instruction sender and the identity information of the target model from the instruction, then determines, according to the identifier of the instruction sender and the identity information of the target model, whether the identifier of the instruction sender exists in a correspondence between identifiers of legitimate users in a pre-stored intelligent contract and the identity information of the target model, and finds information of a storage node corresponding to the identity information of the target model from a correspondence between the pre-stored identity information of the model and information of the storage node according to the identity information of the target model when it is determined that the identifier of the instruction sender exists.
As an example, when it is determined that the identifier of the instruction sender does not exist, the information of the storage node corresponding to the identity information of the target model is not searched, and then step S24 is not executed.
After the management platform determines the information of the at least two storage nodes, step S24 is performed.
S24: and the management platform sends the information of the at least two storage nodes to the instruction sender.
In an actual implementation, S24 may be implemented in such a way that the management platform sends the information of the at least two storage nodes to the instruction sender based on the IPFS protocol.
S25: and the instruction sender receives the information of the at least two storage nodes.
As an embodiment, S25 includes the steps of: A1-A3.
A1: the instruction sender receives the information of the notary node returned by the management platform based on the instruction; wherein the notary node belongs to the management platform.
In this embodiment, the information of the notary node may be an IP address of the notary node.
A2: and the instruction sender sends a node information request to the notary node according to the information of the notary node.
And if the information of the notary node is the IP address of the notary node, an instruction sender sends a node information request to the notary node according to the IP address of the notary node, so that the notary node sends the information of the storage node stored in the node to the instruction sender after receiving the node information request.
A3: and the instruction sender receives the information of the at least two storage nodes returned by the notary node based on the node information request.
S26: and the instruction sender acquires the parameter information of the network model layer stored at each storage node according to the information of each storage node.
If the information of the storage node is the IP address of the storage node, S26 may be implemented in such a manner that the instruction sender sends a request to the storage node according to the IP address of each storage node, so that the storage node sends the parameter information of the network model layer stored by itself to the instruction sender according to the request, and the instruction sender receives the parameter information of the network model layer returned by the storage node based on the request.
As an embodiment, the information of the at least two storage nodes includes: a public key corresponding to each storage node in the at least two storage nodes; s26 includes the steps of: B1-B4.
B1: an instruction sender sends a parameter request to each storage node according to the information of the storage node; wherein the parameter request includes: a public key corresponding to the storage node.
In an actual implementation process, B1 may be implemented as follows, where the instruction sender generates a parameter request for information of each storage node, where the parameter request includes: and then sending the parameter request to each storage node in a management platform in a broadcast manner by using the public key corresponding to the storage node, so that each storage node extracts the public key from the parameter request after receiving the parameter request, then comparing the public key stored by the storage node with the public key in the parameter request, and sending a feedback result to an instruction sender when the public keys are the same, wherein the feedback result comprises: the encrypted key and the encrypted parameter information stored by the device; and when the parameters are different, the parameter request is not responded.
As an implementation manner, after receiving the parameter request, each storage node extracts a public key from the parameter request, then compares the public key stored in itself with the public key in the parameter request, and encrypts the encrypted secret key and the encrypted parameter information stored in itself according to the private key stored in itself to obtain a feedback result when the public keys are the same, and then sends the feedback result to the instruction sender.
B2: the instruction sender receives a feedback result returned by the storage node based on the parameter request; the feedback result comprises: an encrypted key, and encrypted parameter information.
In practical implementation, B2 may be implemented in such a way that the receiver is instructed to receive the feedback result in real time or not.
B3: and the instruction sender decrypts the encrypted secret key by using the corresponding public key to obtain a symmetric secret key.
In an actual implementation process, B3 may be implemented as follows, where the instruction sender extracts the encrypted key from the feedback result, and decrypts the encrypted key with the public key corresponding to the storage node to obtain a symmetric key.
As an embodiment, B3 may be implemented in such a way that the sender is instructed to decrypt the feedback result by using the corresponding public key to obtain an encrypted key and encrypted parameter information, and then decrypt the encrypted key by using the corresponding public key to obtain the symmetric key.
B4: and the instruction sender decrypts the encrypted parameter information by using the symmetric key to obtain the parameter information of the corresponding network model layer.
And the instruction sender decrypts the encrypted parameter information in the feedback result by using the symmetric key to obtain the parameter information of the corresponding network model layer.
As an embodiment, the information of the at least two storage nodes further includes: an identifier of a network model layer corresponding to each of the at least two storage nodes; the method further comprises the steps of: C.
c: and the instruction sending party correspondingly sets the values of all parameters in the initial model of the target model by using the parameter information of the network model layer acquired from each storage node according to the identifier of the network model layer corresponding to each storage node to obtain the target model.
In the implementation process, the instruction of the instruction sender determines, according to the identifier of the network model layer corresponding to each storage node, each initial network model layer in the initial model corresponding to the parameter information of the network model layer acquired from each storage node, and for each initial network model layer in the initial model, sets, by using the parameter information corresponding to the initial network model layer, each parameter in the initial network model layer correspondingly, so as to obtain the target model.
As an embodiment, the management platform further comprises: a notary node; the method further comprises the steps of: D1-D3.
D1: aiming at each storage node in the at least two storage nodes, the notary node continuously receives the Hash value sent by the storage node; and the Hash value is determined by the storage node by using the parameter information of the network model layer stored in the storage node at the current moment.
It can be understood that, for each storage node of the at least two storage nodes, the storage node determines a Hash value based on a preset time interval by using parameter information stored in the storage node at the current time, and sends a public key stored in the storage node and the determined Hash value to the notary node, and the notary node receives the Hash value and the public key and correspondingly stores the Hash value, the public key and the receiving time sent by the storage node for the first time, wherein the Hash value sent by the storage node for the first time is determined by using the parameter information of the network model layer sent by the management platform by the storage node.
As an implementation manner, when the notary node receives the public key and the Hash value sent by the storage node, the notary node stores the corresponding relationship among the public key, the Hash value and the time when the Hash value is received.
D2: the method comprises the steps that when a notary node determines that a Hash value received at the current moment is different from a Hash value stored in advance, parameter information stored on a storage node is determined to be changed; and the pre-stored Hash value is determined by the storage node by using the parameter information of the network model layer of the target model sent by the management platform.
For each storage node in the at least two storage nodes, when a notary receives a Hash value and a public key sent by the storage node at the current moment, the notary node searches the Hash value corresponding to the public key from the Hash value stored by the notary node according to the public key received at the current moment, then compares the Hash value received at the current moment with the corresponding Hash value, determines that the parameter information stored on the storage node changes when the determined values are different, and otherwise determines that the parameter information stored on the storage node does not change.
D3: the notary node generates prompt information; and the prompt information is used for prompting that the parameter information stored on the storage node is tampered.
Wherein, the prompt message includes: the information of the storage node.
As an embodiment, the method further comprises:
and the notary node sends the public key in the first asymmetric key pair randomly generated by the notary node to the at least two storage nodes for storage.
The notary node divides a private key in the first asymmetric key pair into 2n +1 fragments based on a Shamir key sharing principle, wherein the value of n is the number of the at least two storage nodes, the private key in the first asymmetric key pair is used for encrypting the n fragments in the 2n +1 fragments respectively, and each encrypted fragment is sent to each storage node in the at least two storage nodes respectively, and it can be understood that one storage node in the at least two storage nodes only receives one encrypted fragment; for each storage node of the at least two storage nodes, the storage node will return a fragment to the notary node when the storage node and the notary node reach consensus, otherwise, the storage node will not return a fragment to the notary node.
The notary node sends the information of the at least two storage nodes to an instruction sender only when the number of the received fragments is larger than half of the number of the at least two storage nodes, otherwise, the notary node does not send the information of the at least two storage nodes to the instruction sender.
Referring to fig. 3, fig. 3 is a block diagram of an information processing apparatus 300 according to an embodiment of the present disclosure. The block diagram of fig. 3 will be explained, and the apparatus shown comprises:
and an instruction sending unit 310, configured to send a parameter information obtaining instruction of the target model to the management platform.
A receiving unit 320, configured to receive information of at least two storage nodes returned by the management platform based on the instruction; the at least two storage nodes respectively store parameter information of different network model layers in the target model; the at least two storage nodes belong to the management platform.
An obtaining unit 330, configured to obtain parameter information of the network model layer stored at each storage node according to the information of each storage node.
As an embodiment, the receiving unit 320 is specifically configured to receive information of the notary node returned by the management platform based on the instruction; wherein the notary node belongs to the management platform; sending a node information request to the notary node according to the information of the notary node; and receiving the information of the at least two storage nodes returned by the notary node based on the node information request.
As an embodiment, the information of the at least two storage nodes includes: a public key corresponding to each storage node in the at least two storage nodes; the obtaining unit 330 is specifically configured to send a parameter request to each storage node according to information of the storage node; wherein the parameter request includes: a public key corresponding to the storage node; receiving a feedback result returned by the storage node based on the parameter request; the feedback result comprises: an encrypted key, and encrypted parameter information; decrypting the encrypted key by using the corresponding public key to obtain a symmetric key; and decrypting the encrypted parameter information by using the symmetric key to obtain the parameter information of the corresponding network model layer.
As an embodiment, the information of the at least two storage nodes further includes: an identifier of a network model layer corresponding to each of the at least two storage nodes; the device further comprises: and according to the identifier of the network model layer corresponding to each storage node, utilizing the parameter information of the network model layer acquired from each storage node to correspondingly set the value of each parameter in the initial model of the target model, so as to obtain the target model.
According to the management platform provided by the embodiment of the application, the management platform is used for receiving a parameter information acquisition instruction of a target model; determining information of at least two storage nodes according to the instruction; the at least two storage nodes respectively store parameter information of different network model layers in the target model; the at least two storage nodes belong to the management platform; and sending the information of the at least two storage nodes to the instruction sender.
As an embodiment, the management platform further comprises: a notary node; the notary node is used for continuously receiving the Hash value sent by each storage node in the at least two storage nodes; the Hash value is determined by the storage node by using the parameter information of the network model layer stored in the storage node at the current moment; when the Hash value received at the current moment is determined to be different from the Hash value stored in advance, determining that the parameter information stored on the storage node changes; the pre-stored Hash value is determined by the storage node by using the parameter information of the network model layer of the target model sent by the management platform; generating prompt information; and the prompt information is used for prompting that the parameter information stored on the storage node is tampered.
For the process of implementing each function by each functional unit in this embodiment, please refer to the content described in the embodiment shown in fig. 1-2, which is not described herein again.
Referring to fig. 4, fig. 4 is a schematic structural diagram of an electronic device 400 according to an embodiment of the present disclosure, where the electronic device 400 may be an instruction sender or a management platform in the above embodiments, and the electronic device 400 may be a Personal Computer (PC), a tablet computer, a smart phone, a Personal Digital Assistant (PDA), or the like.
The electronic device 400 may include: memory 402, processor 401, communication interface 403, and a communication bus for enabling connection communication of these components.
The Memory 402 is used for storing various data such as a computer program instruction corresponding to the information processing method and apparatus provided in the embodiment of the present application, where the Memory 402 may be, but is not limited to, a Random Access Memory (RAM), a Read Only Memory (ROM), a Programmable Read-Only Memory (PROM), an Erasable Read-Only Memory (EPROM), an electrically Erasable Read-Only Memory (EEPROM), and the like.
The processor 401 is configured to read the corresponding calculation program instructions of the information processing method and apparatus.
The processor 401 may be an integrated circuit chip having signal processing capability. The Processor 401 may be a general-purpose Processor, including a Central Processing Unit (CPU), a Network Processor (NP), and the like; but may also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware components. The various methods, steps, and logic blocks disclosed in the embodiments of the present application may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
A communication interface 403 for receiving or transmitting data.
In addition, a storage medium is provided in an embodiment of the present application, and a computer program is stored in the storage medium, and when the computer program runs on a computer, the computer is caused to execute the method provided in any embodiment of the present application.
In summary, the information processing method, apparatus, management platform, electronic device and storage medium according to embodiments of the present application send a parameter information obtaining instruction of a target model to the management platform, accurately obtain parameter information related to the target model from at least two storage nodes according to information of the storage nodes returned by the management platform, so as to implement sharing of the model, and then store parameter information of different network model layers in the model by using different storage nodes, so that even if an illegal person obtains part of parameter information of the model from one storage node, it is not easy to obtain all parameter information of the model, and therefore, compared with the prior art in which a storage node is directly used to store all parameter information of the model, it is better to prevent the illegal person from directly obtaining all parameter information of the model, the safety of the model is ensured.
In the embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other ways. The apparatus embodiments described above are merely illustrative, and for example, the flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based devices that perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
In addition, functional modules in the embodiments of the present application may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.
Claims (10)
1. An information processing method, characterized in that the method comprises:
sending a parameter information acquisition instruction of the target model to a management platform;
receiving information of at least two storage nodes returned by the management platform based on the instruction; the at least two storage nodes respectively store parameter information of different network model layers in the target model; the at least two storage nodes belong to the management platform;
and acquiring the parameter information of the network model layer stored at each storage node according to the information of each storage node.
2. The method of claim 1, wherein receiving information of at least two storage nodes returned by the management platform based on the instruction comprises:
receiving the notary node information returned by the management platform based on the instruction; wherein the notary node belongs to the management platform;
sending a node information request to the notary node according to the information of the notary node;
and receiving the information of the at least two storage nodes returned by the notary node based on the node information request.
3. The method according to claim 1 or 2, wherein the information of the at least two storage nodes comprises: a public key corresponding to each storage node in the at least two storage nodes; the acquiring parameter information of the network model layer stored at each storage node according to the information of each storage node includes:
sending a parameter request to each storage node according to the information of the storage node; wherein the parameter request includes: a public key corresponding to the storage node;
receiving a feedback result returned by the storage node based on the parameter request; the feedback result comprises: an encrypted key, and encrypted parameter information;
decrypting the encrypted key by using the corresponding public key to obtain a symmetric key;
and decrypting the encrypted parameter information by using the symmetric key to obtain the parameter information of the corresponding network model layer.
4. The method of claim 1, wherein the information of the at least two storage nodes further comprises: an identifier of a network model layer corresponding to each of the at least two storage nodes; the method further comprises the following steps:
and according to the identifier of the network model layer corresponding to each storage node, utilizing the parameter information of the network model layer acquired from each storage node to correspondingly set the value of each parameter in the initial model of the target model, so as to obtain the target model.
5. An information processing method applied to a management platform, the method comprising:
receiving a parameter information acquisition instruction of a target model;
determining information of at least two storage nodes according to the instruction; the at least two storage nodes respectively store parameter information of different network model layers in the target model; the at least two storage nodes belong to the management platform;
and sending the information of the at least two storage nodes to the instruction sender.
6. The method of claim 5, wherein the management platform further comprises: a notary node; the method further comprises the following steps:
aiming at each storage node in the at least two storage nodes, the notary node continuously receives the Hash value sent by the storage node; the Hash value is determined by the storage node by using the parameter information of the network model layer stored in the storage node at the current moment;
the notary node determines that the parameter information stored on the storage node changes when the Hash value received at the current moment is different from the Hash value stored in advance; the pre-stored Hash value is determined by the storage node by using the parameter information of the network model layer of the target model sent by the management platform;
the notary node generates prompt information; and the prompt information is used for prompting that the parameter information stored on the storage node is tampered.
7. An information processing apparatus characterized in that the apparatus comprises:
the instruction sending unit is used for sending a parameter information acquisition instruction of the target model to the management platform;
the receiving unit is used for receiving the information of the at least two storage nodes returned by the management platform based on the instruction; the at least two storage nodes respectively store parameter information of different network model layers in the target model; the at least two storage nodes belong to the management platform;
and the acquisition unit is used for acquiring the parameter information of the network model layer stored at each storage node according to the information of each storage node.
8. The management platform is used for receiving a parameter information acquisition instruction of a target model; determining information of at least two storage nodes according to the instruction; the at least two storage nodes respectively store parameter information of different network model layers in the target model; the at least two storage nodes belong to the management platform; and sending the information of the at least two storage nodes to the instruction sender.
9. An electronic device comprising a memory and a processor, the memory having stored therein computer program instructions that, when read and executed by the processor, perform the method of any of claims 1-6.
10. A computer-readable storage medium having stored thereon computer program instructions which, when read and executed by a computer, perform the method of any one of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011556622.4A CN112738061B (en) | 2020-12-24 | 2020-12-24 | Information processing method, device, management platform, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011556622.4A CN112738061B (en) | 2020-12-24 | 2020-12-24 | Information processing method, device, management platform, electronic equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112738061A true CN112738061A (en) | 2021-04-30 |
| CN112738061B CN112738061B (en) | 2022-06-21 |
Family
ID=75615757
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011556622.4A Active CN112738061B (en) | 2020-12-24 | 2020-12-24 | Information processing method, device, management platform, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112738061B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2023138234A1 (en) * | 2022-01-20 | 2023-07-27 | 北京邮电大学 | Model management method and apparatus, networking architecture, electronic device and storage medium |
Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140280251A1 (en) * | 2013-03-15 | 2014-09-18 | Yahoo! Inc. | Almost online large scale collaborative filtering based recommendation system |
| CN108574699A (en) * | 2018-07-20 | 2018-09-25 | 广东工业大学 | A kind of communication connecting method, system and internet of things equipment system and storage medium |
| CN109194508A (en) * | 2018-08-27 | 2019-01-11 | 联想(北京)有限公司 | Data processing method and device based on block chain |
| CN109472164A (en) * | 2018-09-13 | 2019-03-15 | 远光软件股份有限公司 | The verification method and its network and electronic equipment of contract dataset |
| CN109508485A (en) * | 2018-10-30 | 2019-03-22 | 平安医疗健康管理股份有限公司 | A kind of data processing model dissemination method, device, server and storage medium |
| CN109710583A (en) * | 2018-12-28 | 2019-05-03 | 上海仁童电子科技有限公司 | Parameter automatic update method and device in Simulink model |
| CN109815344A (en) * | 2019-01-29 | 2019-05-28 | 华南师范大学 | Network model training system, method, apparatus and medium based on parameter sharing |
| CN109993287A (en) * | 2017-12-29 | 2019-07-09 | 北京中科寒武纪科技有限公司 | Processing with Neural Network method, computer system and storage medium |
| CN110572253A (en) * | 2019-09-16 | 2019-12-13 | 济南大学 | Method and system for enhancing privacy of federated learning training data |
| US20200092082A1 (en) * | 2018-09-19 | 2020-03-19 | International Business Machines Corporation | Distributed platform for computation and trusted validation |
| CN111027870A (en) * | 2019-12-14 | 2020-04-17 | 支付宝(杭州)信息技术有限公司 | User risk assessment method and device, electronic equipment and storage medium |
| CN111565199A (en) * | 2020-07-14 | 2020-08-21 | 腾讯科技（深圳）有限公司 | Network attack information processing method and device, electronic equipment and storage medium |
| CN111723907A (en) * | 2020-06-11 | 2020-09-29 | 浪潮电子信息产业股份有限公司 | Model training device, method, system and computer readable storage medium |
| CN111970335A (en) * | 2020-07-30 | 2020-11-20 | 腾讯科技（深圳）有限公司 | Information recommendation method and device and storage medium |
| CN112100145A (en) * | 2020-09-02 | 2020-12-18 | 南京三眼精灵信息技术有限公司 | Digital model sharing learning system and method |
-
2020
- 2020-12-24 CN CN202011556622.4A patent/CN112738061B/en active Active
Patent Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140280251A1 (en) * | 2013-03-15 | 2014-09-18 | Yahoo! Inc. | Almost online large scale collaborative filtering based recommendation system |
| CN109993287A (en) * | 2017-12-29 | 2019-07-09 | 北京中科寒武纪科技有限公司 | Processing with Neural Network method, computer system and storage medium |
| CN108574699A (en) * | 2018-07-20 | 2018-09-25 | 广东工业大学 | A kind of communication connecting method, system and internet of things equipment system and storage medium |
| CN109194508A (en) * | 2018-08-27 | 2019-01-11 | 联想(北京)有限公司 | Data processing method and device based on block chain |
| CN109472164A (en) * | 2018-09-13 | 2019-03-15 | 远光软件股份有限公司 | The verification method and its network and electronic equipment of contract dataset |
| US20200092082A1 (en) * | 2018-09-19 | 2020-03-19 | International Business Machines Corporation | Distributed platform for computation and trusted validation |
| CN109508485A (en) * | 2018-10-30 | 2019-03-22 | 平安医疗健康管理股份有限公司 | A kind of data processing model dissemination method, device, server and storage medium |
| CN109710583A (en) * | 2018-12-28 | 2019-05-03 | 上海仁童电子科技有限公司 | Parameter automatic update method and device in Simulink model |
| CN109815344A (en) * | 2019-01-29 | 2019-05-28 | 华南师范大学 | Network model training system, method, apparatus and medium based on parameter sharing |
| CN110572253A (en) * | 2019-09-16 | 2019-12-13 | 济南大学 | Method and system for enhancing privacy of federated learning training data |
| CN111027870A (en) * | 2019-12-14 | 2020-04-17 | 支付宝(杭州)信息技术有限公司 | User risk assessment method and device, electronic equipment and storage medium |
| CN111723907A (en) * | 2020-06-11 | 2020-09-29 | 浪潮电子信息产业股份有限公司 | Model training device, method, system and computer readable storage medium |
| CN111565199A (en) * | 2020-07-14 | 2020-08-21 | 腾讯科技（深圳）有限公司 | Network attack information processing method and device, electronic equipment and storage medium |
| CN111970335A (en) * | 2020-07-30 | 2020-11-20 | 腾讯科技（深圳）有限公司 | Information recommendation method and device and storage medium |
| CN112100145A (en) * | 2020-09-02 | 2020-12-18 | 南京三眼精灵信息技术有限公司 | Digital model sharing learning system and method |
Non-Patent Citations (2)
| Title |
|---|
| THEIN THAN THWIN，SANGSUREE VASUPONGAYYA: "blockchain based secret-data sharing model for personal health record system", 《IEEE》 * |
| 辛海涛: "P2P网络中信任模型的研究", 《中国优秀硕士学术论文信息科技辑》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2023138234A1 (en) * | 2022-01-20 | 2023-07-27 | 北京邮电大学 | Model management method and apparatus, networking architecture, electronic device and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112738061B (en) | 2022-06-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107864115B (en) | Method for user account login verification by using portable terminal | |
| CN101465735B (en) | Network user identification verification method, server and client terminal | |
| CN108512846B (en) | Bidirectional authentication method and device between terminal and server | |
| CN112019541B (en) | Data transmission method and device, computer equipment and storage medium | |
| CN110362357A (en) | A kind of configuration file management method and device of application program | |
| CN110519215B (en) | Data communication method and device | |
| CN113708935B (en) | Internet of things equipment unified authentication method and system based on block chain and PUF | |
| CN109729000B (en) | Instant messaging method and device | |
| CN110866265A (en) | Data storage method, device and storage medium based on block chain | |
| CN113360953B (en) | IPFS and blockchain-based file distributed storage management method and system | |
| CN107517194B (en) | Return source authentication method and device of content distribution network | |
| CN111090386A (en) | Cloud storage method, device and system and computer equipment | |
| CN111367923A (en) | Data processing method, data processing device, node equipment and storage medium | |
| CN111461720A (en) | Identity verification method and device based on block chain, storage medium and electronic equipment | |
| CN112738061B (en) | Information processing method, device, management platform, electronic equipment and storage medium | |
| CN109089265B (en) | Equipment networking management system and method and intelligent equipment | |
| CN111327561A (en) | Authentication method, system, authentication server, and computer-readable storage medium | |
| CN105141624A (en) | Login method, account management server and client system | |
| CN111479265A (en) | Information dissemination method and device, computer equipment and storage medium | |
| CN110232570A (en) | A kind of information monitoring method and device | |
| CN116305013A (en) | Electronic file adding method and device of traceability information, electronic equipment and medium | |
| CN211557285U (en) | Control terminal, signature server and task server | |
| CN114338132A (en) | Secret-free login method, client application, operator server and electronic equipment | |
| CN111083164A (en) | Safety protection method of industrial control system and related equipment | |
| CN114666154B (en) | Device communication method, device, gateway, device, system, medium and product |
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 |