Federal learning method for hiding data source information based on group signature, smart contract and distributed file system

Abstract

The invention discloses a federal learning method for hiding data source information based on group signature, smart contract and distributed file system, and relates to the technical field of computer cryptography. According to the method, the smart contract is utilized to manage each group member, each group member completes model training by utilizing local data to obtain model parameters, the model parameters are encrypted by utilizing a group signature function and then uploaded to the DFS to obtain a DFS address, and then encryption and group signature are performed on the encrypted key and the DFS address through the smart contract, so that the group signature of the encrypted key and the DFS address is realized. And the central server obtains a DFS address and a secret key through decryption, downloads encrypted model parameters from the DFS, obtains model parameters uploaded by each group member through decryption by using the secret key, and obtains an improved global model after aggregation updating by using the model parameters. The privacy of the data provider is effectively protected, so that the data source information of the two parties in the federal learning process is protected.

Description

Technical Field

[0001] The present invention relates to the technical field of computer cryptography, and more specifically to a federated learning method for hiding data source information based on group signatures, smart contracts and a distributed file system (DFS). Background Art

[0002] In today's information age, machine learning, as a core technology for data analysis and pattern recognition, is profoundly changing our work and lifestyle. Its basic principle is to extract valuable information from large amounts of data by training algorithms, thereby enabling computer systems to have the ability to predict and make decisions about new data. From simple linear regression to complex deep learning networks, these algorithms play a vital role in many fields such as medicine, finance, and social media.

[0003] Taking the medical field as an example, machine learning technology has been widely used in disease diagnosis, disease prediction and personalized treatment plan formulation. However, these applications need to process a large amount of sensitive patient data, such as medical records, biometric information and gene sequences. Similarly, in personal application areas, such as smartphone applications and social network services, machine learning technology also plays an important role in personalized recommendations and user behavior analysis, and the data processed often involves users' personal preferences and behavioral habits.

[0004] However, as data privacy and confidentiality issues become increasingly prominent, how to make full use of this data while protecting privacy has become an urgent problem to be solved. Although traditional centralized machine learning methods have the advantage of computational efficiency, they have privacy protection issues when processing sensitive data. This is because this method requires a large amount of data to be centralized in a single data center or server for processing, which increases the risk of data theft or abuse, and may also expose private information during transmission.

[0005] To solve this problem, federated learning has emerged as an emerging machine learning method. It keeps the data on the original data source, such as the user's smartphone or the hospital's server, and then sends the machine learning model to each data source, independently learning on these local devices, and finally sending the learned model update (not the original data) back to the central server. This method effectively reduces the risk of data leakage and also reduces the demand for bandwidth.

[0006] However, although federated learning has made some progress in improving content privacy, it still faces challenges in hiding the address of the data source. For example, in the medical field, even if the patient's specific health data is protected, if the data can be identified as coming from a specific hospital or region, it may still expose the patient's identity or health status. In the field of personal applications, even if the user's specific behavioral data is anonymized, the leakage of the geographic location information of the data source may reveal the user's personal habits and behavior patterns.

[0007] Therefore, it is particularly urgent to develop a federated learning method that can simultaneously protect data content and hide the data source address. Summary of the invention

[0008] In order to overcome the defects and deficiencies in the above-mentioned prior art, the present invention provides a federated learning method for hiding data source information based on group signatures, smart contracts and distributed file systems. The purpose of the present invention is to solve the problem that the prior art cannot hide data source information. In the federated learning method for hiding data sources based on group signatures, smart contracts and distributed file systems proposed by the present invention, during the entire federated learning training process, all interactions between the data holder and the central server do not directly pass through the network, but provide an encryption scheme based on computer cryptography with the help of group signatures, and use smart contracts and DFS to transfer data addresses and data themselves respectively, which protects the source information of the data provider with a slight impact on efficiency, effectively protects the privacy information of the data provider, and protects the data source information of both parties in the process of federated learning.

[0009] In order to solve the above problems existing in the prior art, the present invention is implemented through the following technical solutions.

[0010] The present invention provides a federated learning method for hiding data source information based on group signature, smart contract and distributed system, the method comprising the following steps:

[0011] S1. Create a smart contract: Create a centralized platform for group signature group management, allowing each node to register, and assign group signatures and group private keys to successfully registered nodes, which are used to transmit public keys, encrypted messages and group signatures. The smart contract; wherein the nodes include central server nodes and group member nodes;

[0012] S2. Creation and management of federated learning groups: Create a federated learning group and elect a group administrator. Use the group administrator's computer or server as the oracle of the smart contract. The oracle randomly generates a group signature group, publishes the public key information of the group signature group to the network, and keeps the private key privately. The oracle generates a group certificate and a group private key for each group member who joins the group signature group. After the central server role is registered, its public key used for symmetric encryption is published through the smart contract.

[0013] S3, model training and model parameter upload: The central server transmits the global model to each group member in the group signature group. After using local data to train the federated learning model, the group members obtain the model parameters, encrypt the model parameters with the key to obtain the encrypted parameters, upload the encrypted parameters to DFS, and obtain the DFS address; encrypt the key and DFS address with the public key of the central server to obtain an encrypted message, and use the group certificate and group private key to group sign the encrypted message, and call the smart contract to upload the encrypted message and group signature;

[0014] S4. The central server obtains model parameters and aggregates updates: The central server monitors the smart contract, verifies the group signature using the public key of the group signature group, and obtains the encrypted message; the central server decrypts the encrypted message to obtain the key and DFS address, and uses the DFS address to retrieve the encrypted parameters from DFS, and uses the key to decrypt the encrypted parameters to obtain the model parameters;

[0015] After obtaining sufficient model parameters, the central server uses the set algorithm to perform aggregate updates to improve the global model. The improved global model is sent back to each group member in the group signature group for the next round of local training.

[0016] Further preferably, in step S3, the group members first randomly select a key, symmetrically encrypt the model parameters using the key to obtain the encrypted parameters, and then upload the encrypted parameters to DFS to obtain the DFS address; the group members asymmetrically encrypt the key and the DFS address using the public key of the central server to obtain an encrypted message; the group members use the group certificate and the group private key to group sign the encrypted message, and finally, call the smart contract to upload the encrypted message and the group signature.

[0017] Further preferably, in step S4, the central server uses the public key information of the group signature group to verify the group signature and filters out the encrypted messages from the group signature group; the central server uses its own asymmetric encryption private key to decrypt the encrypted message to obtain the key and DFS address; the central server uses the DFS address to retrieve the encryption parameters from the DFS, and uses the key to decrypt to obtain the model parameters.

[0018] More preferably, after the group members use the local data to train the federated learning model, they obtain the model parameters, and use differential privacy technology to protect the model parameters, so that the characteristics of the data cannot be obtained through the model parameters.

[0019] Further preferably, in step S2, a federated learning group is formed by companies or organizations in the network that hold similar data or attempt to use federated learning to jointly train a model, and members of the federated learning group jointly elect a trusted group administrator.

[0020] More preferably, the group administrator prepares a computer or server as an oracle to implement the interface required by the smart contract and monitor the events of the smart contract; the oracle randomly selects some basic parameters to generate a group signature group, and then publishes the public key information of the group signature group to the network so that everyone can query it, and keeps the private key of the group signature group privately.

[0021] More preferably, when a company or organization wants to join a group signature group of a federated learning group to jointly train a model, the company or organization calls the smart contract to apply to join the federated learning group and passes in its own relevant information.

[0022] More preferably, after the oracle monitors the event of joining the group signature group, it first conducts a qualification review of the newly joined group member. After the review is passed, the oracle uses the private key information of the group signature group to randomly select a parameter to generate a group certificate and a group private key for the newly joined group member, and returns the group certificate and the group private key to the newly joined group member through an encrypted channel.

[0023] Further preferably, when a group member publishes an erroneous model parameter, the group administrator determines the group certificate and the group signature through the erroneous model parameter, and determines the identity of the group member who signed the group signature through the group certificate and the group signature.

[0024] Further preferably, in the step S1, the created smart contract implements the CreateGroup interface, the RegisterNode interface, the RegisterServer interface, the RegisterReturn interface and the UploadMessage interface; wherein,

[0025] The CreateGroup interface declares a group in the smart contract. This interface triggers the GroupCreate event. The group administrator calls this interface to create a group signature group.

[0026] The RegisterNode interface triggers the NodeRegister event, which is listened to by the oracle and passes the node name, node-related information, and node public key of the registered node; this interface is called when a new group member joins the group signature group;

[0027] The RegisterServer interface triggers the ServerRegister event, which is listened to by the oracle and passes the server name, server-related information, and server public key of the registered central server; when a central server role is registered, this interface is called;

[0028] The RegisterReturn interface triggers the ReturnRegister event, which is listened to by the central server and group members, and transmits the registrant's public key, registration results and other information. Specifically, when the oracle listens to the NodeRegister event, the group administrator will conduct a qualification review of the newly joined group member, and then call the RegisterReturn interface to feedback the review result. If the review fails, the review result is failure, and the additional information is the reason for failure. If the review passes, the review result is passed, the additional information is empty, and the group certificate and group private key are passed to the newly joined group member through encrypted information.

[0029] The UploadMessage interface triggers the MessageUpload event, which is monitored by the central server and passes the message together with the group signature of the message; the interface is called when a group member sends a message and signature to the central server.

[0030] Compared with the prior art, the beneficial technical effects brought by the present invention are as follows:

[0031] 1. In the entire federated learning training process, all interactions between the data holder and the central server do not go directly through the network. Instead, an encryption scheme based on computer cryptography is provided with the help of group signatures. At the same time, smart contracts and DFS are used to transmit data addresses and data themselves respectively. The source information of the data provider is protected with a slight impact on efficiency, and the privacy information of the data provider is effectively protected, so that the data source information of both parties is protected in the process of federated learning.

[0032] 2. The present invention can enhance data privacy protection. By using differential privacy technology, group members ensure the privacy of local data when uploading model parameters, which is crucial for federated learning because in this setting, the data does not leave its original storage location.

[0033] 3. The present invention can perform safe and efficient group management. By electing credible group administrators and adopting group signature technology, the identity authentication and anonymity of group members are ensured. This management mechanism helps prevent malicious behavior and can track specific members when necessary.

[0034] 4. The present invention realizes decentralized model parameter storage. By uploading the encrypted model parameters to a distributed file system (DFS), the risk of single point failure is reduced and the robustness of the system is enhanced.

[0035] 5. The present invention optimizes the role of the central server. The central server receives and processes model parameters through smart contracts and DFS, which reduces the complexity of its direct interaction with a large number of members and improves the efficiency of the system.

[0036] 6. The present invention can aggregate and update the model. The central server can verify and aggregate the model parameters from multiple members, and then update the global model. This method maintains the quality of the model while reducing the need for data centralization.

[0037] 7. The model parameters published by each group member in the present invention are traceable. Group administrators can identify members who upload incorrect model parameters when necessary, which helps maintain the overall health and credibility of the federated learning group.

[0038] 8. The present invention makes reasonable use of smart contracts. Smart contracts are used to manage processes such as group joining, model parameter uploading and verification, which increases the transparency and automation of operations and reduces trust dependence and operating costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The foregoing and following detailed description of the present invention will become more apparent when read in conjunction with the following drawings, in which:

[0040] Figure 1 It is a flow chart of the group management function of the technical solution of the present invention.

[0041] Figure 2 It is a schematic diagram of the DFS function of uploading model parameters and the central server obtaining model parameters and aggregating and updating the flow of the technical solution of the present invention.

[0042] Figure 3 It is a flow chart of the model parameter tracing function of the technical solution of the present invention. DETAILED DESCRIPTION

[0043] The technical solution for achieving the purpose of the present invention is further described below through specific embodiments. It should be noted that the technical solution claimed to be protected by the present invention includes but is not limited to the following embodiments.

[0044] Example 1

[0045] As a preferred embodiment of the present invention, this embodiment discloses a federated learning method based on group signatures, smart contracts and distributed systems to hide data source information. The method includes creating a smart contract, creating and managing a federated learning group, training a model and uploading model parameters, and a central server obtaining model parameters and aggregating updates.

[0046] Refer to the instruction manual Figure 1 As shown, create a smart contract, specifically:

[0047] Create a centralized platform for group signature group management, allowing each node to register, and assign group signatures and group private keys to successfully registered nodes, which are used to transmit public keys, encrypted messages and group signature smart contracts; wherein the nodes include central server nodes and group member nodes.

[0048] In this embodiment, a system has one and only one smart contract for all users (group administrators, group members, and central servers) to use.

[0049] As an example, the smart contract implements the CreateGroup interface, the RegisterNode interface, the RegisterServer interface, the RegisterReturn interface and the UploadMessage interface; wherein,

[0050] The CreateGroup interface declares a group in the smart contract. This interface triggers the GroupCreate event. The group administrator calls this interface to create a group signature group.

[0051] The RegisterNode interface triggers the NodeRegister event, which is listened to by the oracle and passes the node name, node-related information, and node public key of the registered node; this interface is called when a new group member joins the group signature group of the federated learning group;

[0052] The RegisterServer interface triggers the ServerRegister event, which is listened to by the oracle and passes the server name, server-related information, and server public key of the registered central server; when a central server role is registered, this interface is called;

[0053] The RegisterReturn interface triggers the ReturnRegister event, which is listened to by the central server and group members, and transmits the registrant's public key, registration results and other information. Specifically, when the oracle listens to the NodeRegister event, the group administrator will conduct a qualification review of the newly joined group member, and then call the RegisterReturn interface to feedback the review result. If the review fails, the review result is failure, and the additional information is the reason for failure. If the review passes, the review result is passed, the additional information is empty, and the group certificate and group private key are passed to the newly joined group member through encrypted information.

[0054] The UploadMessage interface triggers the MessageUpload event, which is monitored by the central server and passes the message together with the group signature of the message; the interface is called when a group member sends a message and signature to the central server.

[0055] As an example, the smart contract only needs to meet the above interfaces and events, and the specific code can be created in Ethereum or Hyperledger Fabric.

[0056] Refer to the instruction manual Figure 1 As shown in the figure, the creation and management of federated learning groups are as follows:

[0057] Create a federated learning group and elect a group administrator. Use the group administrator's computer or server as an oracle. The oracle randomly generates a group signature group, publishes the public key information of the group signature group to the network, and keeps the private key private. The oracle generates a group certificate and a group private key for each group member who joins the group signature group. After the central server role is registered, its public key used for symmetric encryption is made public through a smart contract.

[0058] Refer to the instruction manual Figure 2 As shown, model training and model parameter upload are as follows:

[0059] The central server transmits the global model to each group member in the group signature group. After using local data to train the federated learning model, the group members obtain the model parameters, encrypt the model parameters with the key to obtain the encrypted parameters, upload the encrypted parameters to DFS, and obtain the DFS address; encrypt the key and DFS address with the public key of the central server to obtain an encrypted message, and use the group certificate and group private key to group sign the encrypted message, and call the smart contract to upload the encrypted message and group signature.

[0060] The central server obtains model parameters and aggregates updates, specifically:

[0061] The central server monitors the smart contract, verifies the group signature using the public key of the group signature group, and obtains the encrypted message; the central server decrypts the encrypted message to obtain the key and DFS address, and uses the DFS address to retrieve the encrypted parameters from DFS, and uses the key to decrypt the encrypted parameters to obtain the model parameters;

[0062] After obtaining sufficient model parameters, the central server uses the set algorithm to perform aggregate updates to improve the global model. The improved global model is sent back to each group member in the group signature group for the next round of local training.

[0063] In this embodiment, during the entire federated learning training process, all interactions between the data holder and the central server do not go directly through the network. Instead, an encryption scheme based on computer cryptography is provided with the help of group signatures. At the same time, smart contracts and DFS are used to transfer data addresses and data themselves respectively. Only the central server can access metadata. The source information of the data provider is protected with a slight impact on efficiency, and the privacy information of the data provider is effectively protected, so that the data source information of both parties is protected during the federated learning process.

[0064] In this embodiment, the smart contract is used to transmit the address and encryption information of the model parameters encrypted by the central server public key, ensuring that only the central node can access the metadata; at the same time, it provides a centralized platform for group signature management, allowing each node to safely register its own signature. There is only one smart contract in a system for all users to use.

[0065] Example 2

[0066] As another preferred embodiment of the present invention, this embodiment further supplements and elaborates on the technical solution of the present invention on the basis of the above-mentioned embodiment 1.

[0067] In this embodiment, model training and model parameter uploading are further refined as follows: group members first randomly select a key, symmetrically encrypt the model parameters using the key to obtain encrypted parameters, and then upload the encrypted parameters to DFS to obtain the DFS address; group members use the public key of the central server to asymmetrically encrypt the key and the DFS address to obtain an encrypted message; group members use the group certificate and group private key to group sign the encrypted message, and finally, call the smart contract to upload the encrypted message and group signature.

[0068] In this embodiment, the central server obtains model parameters and aggregates updates, which are further implemented as follows: the central server uses the public key information of the group signature group to verify the group signature and filter out the encrypted messages from the group signature group; the central server uses its own asymmetric encryption private key to decrypt the encrypted message to obtain the key and DFS address; the central server uses the DFS address to retrieve the encrypted parameters from the DFS, and uses the key to decrypt to obtain the model parameters.

[0069] In the above implementation scheme, group members obtain model parameters after training the federated learning model using local data. Differential privacy technology is used to protect the privacy of model parameters, so that the characteristics of the data cannot be obtained through the model parameters. By using differential privacy technology, group members ensure the privacy of local data when uploading model parameters, which is crucial for federated learning because in this setting, the data does not leave its original storage location.

[0070] Example 3

[0071] As another preferred embodiment of the present invention, this embodiment further supplements and elaborates on the technical solution of the present invention on the basis of the above-mentioned embodiment 1 or embodiment 2.

[0072] In this embodiment, the creation and management of the federated learning group is specifically implemented as follows:

[0073] Companies or organizations in the network that hold similar data or attempt to use federated learning to jointly train a model form a federated learning group, and members of the federated learning group jointly elect a trusted group administrator.

[0074] The group administrator prepares a computer or server as an oracle to implement the interface required by the smart contract and monitor the events of the smart contract; the oracle randomly selects some basic parameters to generate a group signature group, and then publishes the public key information of the group signature group to the network so that everyone can query it, and keeps the private key of the group signature group privately.

[0075] When a company or organization wants to join a group signature group of a federated learning group to jointly train a model, the company or organization calls the smart contract to apply to join the group signature group and passes in its own relevant information.

[0076] After the oracle monitors the event of joining the group signature group, it first conducts a qualification review of the newly joined group member. After the review is passed, the oracle uses the private key information of the group signature group to randomly select a parameter to generate a group certificate and group private key for the newly joined group member, and returns the group certificate and group private key to the newly joined group member through an encrypted channel.

[0077] Furthermore, federated learning usually requires the use of massive amounts of data from many nodes, so anomalies are inevitable. If the wrong model parameters are published due to unintentional or malicious operations, the group administrator can, under certain conditions, determine the group certificate and group signature through the wrong model parameters, determine the identity of the group member who signed the group signature through the group certificate and group signature, and take corresponding measures (such as warnings, removing members, etc.) to maintain the overall reputation of the group.

[0078] Example 4

[0079] As another preferred embodiment of the present invention, this embodiment provides an encryption scheme based on computer cryptography with the help of group signatures, and uses the address and data itself of the data to be transmitted using smart contracts and DFS respectively, thereby protecting the source information of the data provider with a slight impact on efficiency, and effectively protecting the privacy information of the data provider.

[0080] In this embodiment, a federated learning method based on group signatures, smart contracts and distributed file systems that hides data source information is used. During the entire federated learning training process, all interactions between the data holder and the central server do not go directly through the network, but instead pass the data address and the data itself through smart contracts and DFS respectively. In this process, the data source information of both parties is protected.

[0081] Specifically include the following:

[0082] S1. Create a smart contract: Create a centralized platform for group signature group management, allowing each node to register, and assign group signatures and group private keys to successfully registered nodes, which are used to transmit public keys, encrypted messages and group signatures. The smart contract; wherein the nodes include central server nodes and group member nodes;

[0083] The created smart contract implements the CreateGroup interface, RegisterNode interface, RegisterServer interface, RegisterReturn interface and UploadMessage interface;

[0084] S2. Creation and management of federated learning groups

[0085] 1) Prepare to create a group. A federated learning group N is formed by companies or organizations n that hold similar data in the network or try to use federated learning to jointly train a model. Group members in the federated learning group jointly elect a trusted group administrator A;

[0086] 2) Create a group signature group. Group administrator A prepares a computer (or server) as the oracle O of the smart contract, implements the interface required by the smart contract, and monitors the events of the smart contract.

[0087] The oracle O randomly selects some basic parameters b to generate a group signature group G, and calls the CreateGroup interface of the smart contract, passing in the group name and group description to form a GroupCreate event.

[0088] Then the public key information gpk of this group is published on the network so that everyone can query it, and the private key secret gmsk is kept privately to ensure that no one can get this private key.

[0089] When a group member wants to join the group signature group as a central server role, the RegisterServer interface of the smart contract is called to trigger the ServerRegister event. After the oracle O of the smart contract listens to the ServerRegister event, it will review the information of the group member registered as the central server role. After the review is passed, the node name, node-related information and node public key of the group member's registered node will be managed, and the public key pubKey used by the central server for symmetric encryption will be made public through the smart contract.

[0090] 3) Join the group. When a company or organization i wants to join a group signature group of federated learning to jointly train a model, the company or organization calls the smart contract method to apply to join the group signature group and passes in its own relevant information.

[0091] When a user wants to join the group signature group as a data holder, the RegisterNode interface of the smart contract is called to trigger the NodeRegister event. After the oracle o listens to the NodeRegister event of joining the group, it first conducts a qualification review of the company or organization. After the review is completed, the oracle calls the RegisterReturn interface of the smart contract, triggers the ReturnRegister event, and returns the review result to member i; if the review fails, the review result is failed, and the additional information is the reason for failure; if the review passes, the review result is passed, and the additional information is empty. The oracle O uses the group private key gmsk and selects the random parameter r to generate the group certificate σ and group private key x for this group member i i , and group certificate σ and group private key x i The group certificates σ are used for signature and are different from each other.

[0092]

[0093] S3, model training and model parameter upload

[0094] 1) Model training. The central server C transmits the model m to each member i. After member i uses the local data data to train the federated learning model, it obtains the model parameters θ. Group member i can use differential privacy and other means to ensure that the data features cannot be obtained through the model parameters θ, thereby protecting data privacy;

[0095]

[0096] 2) Model parameter upload to DFS. Group member i first randomly selects a key key, symmetric encrypts the model parameter using the key key to obtain the encrypted parameter θ', and then uploads the encrypted parameter θ' to DFS to obtain the address addr;

[0097] Subsequently, member i uses the public key pubKey of the central server C to asymmetrically encrypt the key key and address addr to obtain the encrypted message m′. Member i then uses the group certificate σ and the group private key x i Perform group signing on the encrypted message m′ to obtain the signature sig, and finally call the smart contract to upload the encrypted message m′ and signature sig;

[0098]

[0099] When member i sends the encrypted message m′ and signature sig to the central server, the UploadMessage interface of the smart contract is called to trigger the MessageUpload event.

[0100] S4. The central server obtains model parameters and aggregates updates

[0101] 1) Get model parameters. The central server C monitors the smart contract. After the central server monitors the MessageUpload event, it uses the group public key gpk to verify the signature sig and filter out the encrypted message m′ from this group;

[0102] Subsequently, the central server C uses its own asymmetric encryption private key secKey to decrypt the encrypted message m′ and obtain the key key and address addr. The central server C uses the address addr to retrieve the encrypted parameter θ' from DFS and decrypts it using the key key to obtain the model parameter θ;

[0103]

[0104] 2) Aggregate update model. After obtaining enough model parameters θ, the central server C uses a specific algorithm (e.g., weighted averaging, etc.) to perform aggregate updates to improve the global model m, thereby achieving an effect close to that of traditional centralized machine learning. The improved global model m′ is then sent back to group member i for the next round of local training;

[0105] Model parameter traceability: Federated learning usually requires the use of massive data from many nodes, so anomalies are inevitable. If an incorrect model parameter θ is released due to unintentional or malicious operations, group administrator A can open (σ, θ, s) under certain conditions (such as multiple incorrect releases, member requests, etc.) to confirm the identity of group member i who signed the signature. Appropriate measures (such as warnings, removing members, etc.) will be taken accordingly to maintain the overall reputation of the group.

Claims

1. A federated learning method based on group signatures, smart contracts and distributed systems with hidden data source information, characterized by: The method comprises the following steps: S1. Create a smart contract: Create a centralized platform for group signature group management, allowing each node to register, and assign group signatures and group private keys to successfully registered nodes, which are used to transmit public keys, encrypted messages and group signatures. The smart contract; wherein the nodes include central server nodes and group member nodes; S2. Creation and management of federated learning groups: Create a federated learning group and elect a group administrator. Use the group administrator's computer or server as an oracle. The oracle randomly generates a group signature group, publishes the public key information of the group signature group to the network, and keeps the private key private. The oracle generates a group certificate and a group private key for each group member who joins the group signature group. After the central server role is registered, its public key used for symmetric encryption is published through a smart contract. S3, model training and model parameter upload: The central server transmits the global model to each group member in the group signature group. After using local data to train the federated learning model, the group members obtain the model parameters, encrypt the model parameters with the key to obtain the encrypted parameters, upload the encrypted parameters to DFS, and obtain the DFS address; encrypt the key and DFS address with the public key of the central server to obtain an encrypted message, and use the group certificate and group private key to group sign the encrypted message, and call the smart contract to upload the encrypted message and group signature; S4. The central server obtains model parameters and aggregates updates: The central server monitors the smart contract, verifies the group signature using the public key of the group signature group, and obtains the encrypted message; the central server decrypts the encrypted message to obtain the key and DFS address, and uses the DFS address to retrieve the encrypted parameters from DFS, and uses the key to decrypt the encrypted parameters to obtain the model parameters; After obtaining sufficient model parameters, the central server uses the set algorithm to perform aggregate updates to improve the global model. The improved global model is sent back to each group member in the group signature group for the next round of local training.

2. The method for federated learning of hidden data source information based on group signature, smart contract and distributed system as claimed in claim 1, characterized in that: In step S3, the group members first randomly select a key, symmetric encrypt the model parameters using the key to obtain the encrypted parameters, and then upload the encrypted parameters to DFS to obtain the DFS address; the group members asymmetric encrypt the key and DFS address using the public key of the central server to obtain the encrypted message; Group members use the group certificate and group private key to sign the encrypted message. Finally, they call the smart contract to upload the encrypted message and group signature.

3. The method for federated learning of hidden data source information based on group signature, smart contract and distributed system as claimed in claim 2, characterized in that: In step S4, the central server uses the public key information of the group signature group to verify the group signature and filters out the encrypted messages from the group signature group; the central server uses its own asymmetric encryption private key to decrypt the encrypted message to obtain the key and DFS address; the central server uses the DFS address to retrieve the encryption parameters from the DFS and decrypts them using the key to obtain the model parameters.

4. The method for federated learning based on group signature, smart contract and distributed system with hidden data source information according to any one of claims 1 to 3, characterized in that: After using local data to train the federated learning model, group members obtain model parameters and use differential privacy technology to protect the privacy of the model parameters, making it impossible to obtain data features through the model parameters.

5. The method for federated learning of hidden data source information based on group signature, smart contract and distributed system according to any one of claims 1 to 3, characterized in that: In step S2, companies or organizations in the network that hold similar data or attempt to use federated learning to jointly train a model form a federated learning group, and members of the federated learning group jointly elect a trusted group administrator.

6. The method for federated learning based on group signature, smart contract and distributed system with hidden data source information according to any one of claims 1 to 3, characterized in that: The group administrator prepares a computer or server as an oracle to implement the interface required by the smart contract and monitor the events of the smart contract; the oracle randomly selects some basic parameters to generate a group signature group, and then publishes the public key information of the group signature group to the network so that everyone can query it, and keeps the private key of the group signature group privately.

7. The method for federated learning of hidden data source information based on group signature, smart contract and distributed system as claimed in claim 6, characterized in that: When a company or organization wants to join a federated learning group to jointly train a model, the company or organization calls the smart contract to apply to join the group signature group in the federated learning group and passes in its own relevant information.

8. The method for federated learning based on group signature, smart contract and distributed system for hiding data source information as claimed in claim 7, characterized in that: After the oracle monitors the event of joining the group signature group, it first conducts a qualification review of the newly joined group member. After the review is passed, the oracle uses the private key information of the group signature group to randomly select a parameter to generate a group certificate and group private key for the newly joined group member, and returns the group certificate and group private key to the newly joined group member through an encrypted channel.

9. The method for federated learning based on group signature, smart contract and distributed system with hidden data source information according to any one of claims 1 to 3, characterized in that: When a group member publishes an erroneous model parameter, the group administrator determines the group certificate and the group signature through the erroneous model parameter, and determines the identity of the group member who signed the group signature through the group certificate and the group signature.

10. The method for federated learning based on group signature, smart contract and hidden data source information of distributed system according to any one of claims 1 to 3, characterized in that: In the step S1, the created smart contract implements the CreateGroup interface, RegisterNode interface, RegisterServer interface, RegisterReturn interface and UploadMessage interface; wherein, The CreateGroup interface declares a group in the smart contract. This interface triggers the GroupCreate event. The group administrator calls this interface to create a group signature group. The RegisterNode interface triggers the NodeRegister event, which is listened to by the oracle and passes the node name, node-related information, and node public key of the registered node; this interface is called when a new group member joins the group signature group; The RegisterServer interface triggers the ServerRegister event, which is listened to by the oracle and passes the server name, server-related information, and server public key of the registered central server; when a central server role is registered, this interface is called; The RegisterReturn interface triggers the ReturnRegister event, which is monitored by the central server and group members and transmits the registrant's public key, registration results, and other information; The UploadMessage interface triggers the MessageUpload event, which is monitored by the central server and transmits the encrypted message together with the group signature of the encrypted message; the interface is called when a group member sends an encrypted message and group signature to the central server.

Images