Anonymous authentication and security aggregation method oriented to medical data federated learning

Through anonymous authentication and encryption technology, the problems of privacy leakage and security risks in medical data sharing in traditional authentication methods are solved, the anonymization and data security of federated learning are achieved, and the compliance and cooperation of medical data are promoted.

CN119051931AActive Publication Date: 2024-11-29CHONGQING UNIV OF POSTS & TELECOMM

Patent Information

Application Number
CN202411128942.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-16
Publication Date
2024-11-29
Estimated Expiration
2044-08-16

AI Technical Summary

Technical Problem

In medical scenarios, traditional authentication methods can lead to privacy breaches and security risks, especially in data sharing and federated learning. How to ensure identity authenticity and data security is a challenge.

Method used

Using anonymous authentication method, pseudo-identity and long-term keys are generated through trusted institutions and aggregation servers, combined with hash functions and elliptic curve cryptography, the anonymization and data encryption of federated learning participants are realized, ensuring the security of identity authentication and data transmission. sex.

Benefits of technology

Effectively protect patient privacy, reduce attack risks, enhance the security and compliance of medical data, promote medical data sharing and cooperation, and improve the overall security of the system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119051931A_ABST
    Figure CN119051931A_ABST
Patent Text Reader

Abstract

The invention belongs to the technical field of privacy protection, and particularly relates to a medical data federated learning-oriented anonymous authentication and security aggregation method, which comprises the following steps of: initializing a system; registering a federal learning participant Ui and generating a pseudo identity; generating a key; model training; generating a random vector; signature authentication; generating a session key; the aggregation server carries out security aggregation; through the whole system, the non-linkability is realized; the security and confidentiality of the privacy of federated learning participants in the data processing process are greatly improved; privacy protection of gradient parameters: the privacy of personal information of federated learning participants is protected while the data security is improved; the possibility that an attacker obtains sensitive information and data is effectively reduced, and the attack resistance of the system is enhanced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of privacy protection, and specifically relates to an anonymous authentication and security aggregation method for medical data federated learning. Background Art

[0002] Machine learning is to generate a machine learning model through training data computers, and use the generated model and some input data to predict the corresponding results. At present, the demand for high-quality data in machine learning is increasing, and it is necessary to aggregate data distributed in different devices and institutions. The continuous increase in data volume and model complexity makes it difficult for a single node to support model training. Distributed machine learning can achieve a perfect match between training speed and training accuracy by dividing training data and allocating training tasks.

[0003] With the development of big data and artificial intelligence, large-scale data is usually generated by multiple parties and stored in a geographically distributed manner. A traditional and simple distributed machine learning method is to first collect data and store it in a central server and process the data. The prerequisite for this method is that the central server is trustworthy. However, if a large amount of data is sent from multiple devices to a central server, it will generate huge communication overhead, and the source and authenticity of the data are difficult to guarantee. At the same time, data may be missing due to single point failure. It will bring significant risks to the machine learning effect and decision-making results. In medical scenarios, privacy and security have always been the focus of attention. The protection of medical data is crucial to maintaining patient privacy and complying with laws and regulations. However, data sharing and information exchange in medical scenarios usually require cooperation between medical institutions, patients and other relevant parties. The parameter transmission in federated learning in medical scenarios also needs to ensure security and credibility. Traditional authentication methods, such as usernames and passwords, digital certificates, etc., although they can ensure the authenticity of identity to a certain extent, there are some challenges in medical scenarios. First, traditional identity authentication methods may require the provision of identity information, which may involve the risk of privacy leakage. Second, the security of medical data needs to be highly guaranteed, and traditional authentication methods may be at risk of being attacked or misappropriated. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides an anonymous authentication and security aggregation method for medical data federated learning, comprising the following steps:

[0005] Step 1: System initialization;

[0006] Initialize system parameters and keys according to the trusted authority TA, randomly select a set of secure hash functions, and generate a master public key and a master private key;

[0007] Step 2: Federated Learning Participants Ui Registration and pseudo-identity generation;

[0008] According to the trusted institution TA, the federated learning participants U i Randomly select a temporary private key and generate a temporary public key for the pseudo-identity, which is sent to the trusted authority TA. The trusted authority TA verifies the validity of the pseudo-identity and generates complete pseudo-identity information;

[0009] Step 3: Generate key;

[0010] According to the trusted institution TA and the aggregation server Server, the pseudo identity information and the real identity information are respectively input into the trusted institution TA and the aggregation server Server to generate a long-term key;

[0011] Step 4: Model training;

[0012] The aggregation server Server initializes the global model to obtain the global model parameters and distributes the global model parameters to all hospitals’ federated learning participants U i ;

[0013] Step 5: Generate random vectors;

[0014] Generate a random vector based on its own pseudo-identity information and perform a mask operation on the locally trained global model parameters;

[0015] Step 6: Signature authentication;

[0016] The signature is authenticated based on the local gradient model, and then the signature authentication information is sent to the aggregation server Server, which verifies the validity of the signature and sends it to the federated learning participant U i Conduct certification;

[0017] Step 7: Session key generation;

[0018] Federated Learning Participants i The aggregation server generates a session key based on the pseudo-identity information for encryption and decryption of global model parameter transmission;

[0019] Step 8: The aggregation server performs security aggregation;

[0020] The aggregation server securely aggregates the global model parameter update values ​​uploaded by the hospital to obtain the federated average model parameters, and sends the federated average model parameters to the federated learning participants U of each hospital using the session key. i .

[0021] Preferably, the system initialization specifically includes:

[0022] The trusted authority TA is given a finite cyclic additive group G on the elliptic curve with a generator of P and an order of q, and a security parameter k;

[0023] Step 11: The trusted institution TA randomly selects s∈Z q * As the master private key, the master public key is P pub =sP, where: Z q * represents a finite field, that is, the set of integers from 0 to q-1.

[0024] Step 12: The trusted institution TA selects 7 secure hash functions, where the output results of H1 and H2 are the same length as the real identity RID of the participant:

[0025] H1: G×G→{0,1} |RID|

[0026] H2: G×G×{0,1}*×G→{0,1} |RID|

[0027] H3: {0,1} * ×G→Z q *

[0028] H4: {0,1} * ×G×G→Z q *

[0029] H5: {0,1} * ×{0,1} * ×G×G×G×{0,1} * ×{0,1} * →Z q *

[0030] H6: {0,1} * ×{0,1} * ×G×G×G×{0,1} * →Z q *

[0031] H7: {0,1} * ×{0,1} * ×G→Z q *

[0032] Where: H1 represents the first hash function, H2 represents the second hash function, H3 represents the third hash function, H4 represents the fourth hash function, H5 represents the fifth hash function, H6 represents the sixth hash function, H7 represents the seventh hash function, |RID| represents the length of the real identity RID of the participant, → represents the mapping, {0,1} * Represents a string of any length composed of 0 and 1.

[0033] Step 13: The trusted institution TA publishes system parameters param = {q, P, G, P pub ,H1,H2,H3,H4,H5,H6,H7}.

[0034] Preferably, the federated learning participant U i Registration and pseudo-identity generation specifically include:

[0035] Federated Learning Participants i Generate partial pseudo-identity and register with trusted institution TA as federated learning participant U i Generate a pseudo-identity.

[0036] Preferably, the federated learning participant U i Generate some pseudo identities and register them, including:

[0037] Step 21: Federated Learning Participants U i Randomly select a temporary private key x i ∈Z q * , calculate the temporary public key as X i =x i P;

[0038] Step 22: Federated Learning Participants U i Choose a random number y i ∈Z q * , calculate the temporary public key Y of the pseudo-identity i =y i P;

[0039] Step 23: Federated Learning Participants U i Calculate using real identity RID i Calculate A i =RID i ⊕H1(y i P pub ,Y i ), and the federated learning participant U i Send(RID i ,X i ,Y i ,A i ) to the trusted institution TA;

[0040] Among them, Z q * represents a finite field, a set of integers from 0 to q-1, P represents the generator on the elliptic curve given by the trusted authority TA, H1 represents the first hash function, P pub Represents the master public key, A i Indicates RID i Signature.

[0041] Preferably, the trusted institution TA is a federated learning participant U i Generating a pseudo identity specifically includes:

[0042] Step 24: The trusted authority TA calculates the RID i '=A i ⊕H1(sY i ,Y i ), verify the calculated RID i 'Is it equal to the real identity RID sent? i , if they are equal, continue, real identity RID i has not been tampered with, otherwise the registration application will be rejected;

[0043] Step 25: The trusted institution TA calculates the federated learning participant U i Pseudo-identity PID i =RID i ⊕H2(sY i ,Y i ,PIV i ,P pub );

[0044] Step 26: Federated Learning Participant U i The complete pseudo identity is CPID i =(PID i ,PIV i ), where PIV i It's U i The validity period of the pseudo identity, after which the federated learning participant U i Need to re-register to generate a pseudo identity, the trusted institution TA sends CPID i To federated learning participants i ;

[0045] Step 27: The trusted institution TA announces the federated learning participants U i List L = {X i ,PID i ,PIV i};

[0046] Among them, RID i' represents the real identity calculated by the trusted institution TA, A i Indicates RID i The signature of H1 represents the first hash function, H2 represents the second hash function, and Y i represents a temporary public key, s represents the master private key, PIV i Indicates U i The validity period of the pseudo-identity, P pub Represents the master public key, X i Represents a temporary private key.

[0047] Preferably, the key generation specifically includes:

[0048] Step 31: The trusted institution TA selects a random number r i ∈Z q * , calculate U i The public key is R i =r i P, the trusted institution TA calculates B i =H3(PID i ,R i ), the trusted institution TA calculates C i =H4(PID i ,sX i ,R i ), the private key is s i =r i +B i s, sent by the trusted institution TA (s i ,R i ,B i ,C i ) to federated learning participants U i ;

[0049] Step 32: The aggregation server selects a random number x s ∈Z q * , calculate the temporary public key X s =x s P, X s Sent to the trusted institution TA, which selects a random number r s ∈Z q * , calculate the aggregation server Server public key is R s =r s P, the trusted institution TA calculates B s =H3(RID s ,R s ), the trusted institution TA calculates C s =H4(RID s ,sXs ,R s ), the private key is s s =r s +B s s, sent by the trusted institution TA (s s ,R s ,B s ,C s ) to the aggregation server Server;

[0050] Step 33: Federated Learning Participant U i Calculate C i '=H4(PID i ,x i P pub ,R i ), verify the calculated C i 'Is it equal to the sent C i , if they are equal, the key is accepted, otherwise it is rejected. The federated learning participant U i Get PID i The long-term key (s i ,R i );

[0051] Step 34: Aggregation server Server calculates C s '=H4(RID s ,x s P pub ,R s ), verify the calculated C s 'Is it equal to the sent C s If they are equal, the key is accepted, otherwise it is rejected. The aggregation server obtains the long-term key (s s ,R s );

[0052] Among them, Z q * represents a finite field, a set of integers from 0 to q-1, P represents the generator on the elliptic curve given by the trusted authority TA, H3 represents the third hash function, PID i Indicates U i The pseudo identity of H4 is the fourth hash function, s is the master private key, and X i Represents a temporary private key, P pub represents the master public key, B i Indicates U i Part of the private key, C i Indicates U i The signature of the public key, B s Represents part of the aggregation server's Server private key, RID sRepresents the true identity of the aggregation server Server, C i ’ Represents the public key signature of Ui calculation, C s Represents the public key signature of the aggregation server Server, C s ’ Represents the public key signature calculated by the aggregation server Server.

[0053] Preferably, the model training specifically includes: The aggregation server server selects users participating in federated learning from the global model (with all model parameters set to zero) and sets the sequence index id i , and publicly discloses the user list L2 = {X i , PID i , PIV i , id i} and distributes the model to all federated learning participants U i . U i Performs training on the local data and the model sent by the server locally to obtain w i ;

[0054] Among them, X i Represents the temporary private key, PID i Represents the pseudo-identity of U i , PIV i Represents the validity period of the pseudo-identity of U i .

[0055] Preferably, the random vector generation specifically includes:

[0056] The federated learning participant U i Obtains the participant list L2 = {X i , PID i , PIV i , id i}, verifies whether the validity period of the user's pseudo-identity has expired. For valid users U j Calculates the seed x i of the pseudo-random number generator PRG for the federated learning participant U i,j = x i X j , and expands this value into a random vector: p i,j = Δ i,j ·PRG(x i,j ), where, when i > j, Δ i,j = 1; when i < j, Δ i,j = -1, U i Updates the model W i and then adds the random vector to W i to obtain: W i = w i + ∑ i≠jp i,j ; where, when i > j, Δ i,j = 1; when i < j, Δ i,j = -1, U i After updating the model W i to W i by adding a random vector, we get: W i = w i + ∑ i≠j p i,j , and X i represents the temporary private key, PID i represents the pseudo-identity of U i , PIV i represents the validity period of the pseudo-identity of U i , and id i represents the sequence index.

[0057] Preferably, the signature authentication specifically includes:

[0058] Step 61. The federated learning participant U i performs signature authentication on the gradient:

[0059] The federated learning participant U i selects a random number e i ∈ Z q * , E i = e i P, the federated learning participant U i selects a random number f i ∈ Z q * , F i = f i P, the federated learning participant U i calculates k i = H5(RID s , PID i , R i , E i , F i , T i , W i ), u i = e i + k i s i , and the signature is σ i = (k i , u i , R i ), and the federated learning participant Ui sends (W i , k i , u i , R i , F i , Ti ,CPID i ) to the aggregation server Server;

[0060] Step 62: Aggregation Server Verification:

[0061] Aggregation Server Check PIV i Whether it is expired, get the current timestamp T now , calculate |T now -T i | value, if |T now -T i |<Δt, the signature is accepted, otherwise authentication ends. Server calculates B i '=H3(PID i ,R i ), the aggregation server calculates E i '=u i P i (R i +B i 'P pub ), the aggregation server verifies h i =H5(RID s ,PID i ,R i ,E i ',F i ,T i ,W i ) is true, if true then continue, if not true then refuse to accept the message;

[0062] Step 63: Aggregation Server Authentication:

[0063] The aggregation server Server selects a random number e s ∈Z q * , E s =e s P, the aggregation server Server selects a random number f s ∈Z q * , F s =f s P, the aggregation server calculates k s =H6(RID s ,PID i ,R s ,E s ,F s ,T s ),u s =e s +ks s s , the aggregation server sends (k s ,u s ,F s ,R s ,RID s ,T s ) to federated learning participants U i ;

[0064] Step 64: Federated Learning Participant U i Certification:

[0065] Federated Learning Participants i Get the current timestamp T now , calculate |T now -T s | value, if |T now -T s |<Δt, then continue; otherwise, end the authentication. i Calculate B s '=H3(RID s ,R s ), federated learning participants U i Calculate E s '=u s P s (R s +B s 'P pub ), federated learning participants U i Verify i =H6(RID s ,PID i ,R s ,E s ',F s ,T s ) is established, if established, continue;

[0066] Among them, Z q * represents a finite field, a set of integers from 0 to q-1, P represents the generator on the elliptic curve given by the trusted authority TA, H5 represents the fifth hash function, PID i Indicates U i The pseudo-identity of R i Indicates U i Public key, CPID i Represents the federated learning participant U i Complete pseudo identity, H3 represents the third hash function, H6 represents the sixth hash function, k i Indicates partial signature, RID s Indicates the real identity of the aggregation server Server, Ti Indicates U i The current time calculated, timestamp; Δt represents the set time difference, B i 'Server calculation part W i Signed, E. i 'Indicates the part W calculated by the Server i Signature, h i Indicates U i Calculated part W i Signature.

[0067] Preferably, the session key generation specifically includes:

[0068] Step 71: Federated Learning Participant U i Session key generation:

[0069] sk i,s =H7(RID s ,PID i ,f i F s );

[0070] Step 72: Generate the aggregation server session key:

[0071] sk s,i =H7(RID s ,PID i ,fsF i );

[0072] Among them, H7 represents the seventh hash function, PID i Indicates U i The pseudo-identity, f i and f s Represents the federated learning participant U i Choose a random number, F s and F i Indicates that the aggregation server Server selects a random number f s The product of .

[0073] Beneficial effects of the present invention:

[0074] The present invention provides an anonymous authentication and security aggregation method for medical data federated learning. In medical scenarios, anonymous authentication is used to enhance the privacy security of federated learning security aggregation. It can not only effectively protect the privacy of patients and meet regulatory compliance requirements, but also promote medical data sharing and cooperation, and improve medical information security. The present invention introduces a pseudo identity with an expiration date to replace the real identity for authentication, and realizes the anonymization of the identities of federated learning participants while ensuring security;

[0075] Compared with the prior art, the present invention has the following advantages and positive effects:

[0076] 1. Achieve unlinkability: This invention uses a pseudo-identity with an expiration date instead of a real identity for authentication, effectively preventing the possibility of attackers linking different authentication entities or behaviors together through existing information, thereby achieving unlinkability. This improvement greatly improves the security and confidentiality of the privacy of federated learning participants during data processing.

[0077] 2. Privacy protection of gradient parameters: This invention uses random number masking technology to shield gradient parameters, effectively preventing attackers from inferring the sender of parameters by analyzing model parameters. Without leaking the original gradient parameters, the privacy protection of gradient parameters is achieved. This innovative measure not only improves data security, but also protects the privacy of personal information of federated learning participants.

[0078] 3. Improve system security: The improved scheme of the present invention can not only effectively protect the privacy of federated learning participants, but also improve the overall security of the system. By introducing pseudo-identities with validity periods and random number masking technology, it effectively reduces the possibility of attackers obtaining sensitive information and data, and enhances the system's ability to resist attacks. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] Attached Figure 1 It is the basic principle diagram of the present invention;

[0080] Attached Figure 2 This is a system model diagram of the present invention;

[0081] Attached Figure 3 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0082] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0083] Various embodiments of the present invention may be presented in the form of a range. It should be understood that the description in the form of a range is only for convenience and simplicity and should not be understood as a rigid limitation on the scope of the present invention; therefore, it should be considered that the range description has specifically disclosed all possible sub-ranges and single values ​​within the range. For example, it should be considered that the range description from 1 to 6 has specifically disclosed sub-ranges, such as from 1 to 3, from 1 to 4, from 1 to 5, from 2 to 4, from 2 to 6, from 3 to 6, etc., as well as single numbers within the range, such as 1, 2, 3, 4, 5 and 6, regardless of the range. In addition, whenever a numerical range is indicated in the present invention, it is intended to include any quoted numbers (fractions or integers) within the indicated range. Unless otherwise specified, the various raw materials, reagents, instruments and equipment used in the present invention can be purchased from the market or can be prepared by existing equipment.

[0084] In the present invention, in the absence of any contrary description, the directional words used, such as "upper" and "lower", are specifically the directions of the drawings in the accompanying drawings. In addition, in the present invention, the terms "include", "comprise", etc. refer to "including but not limited to". In the present invention, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. In the present invention, "and / or" describes the association relationship of associated objects, indicating that three relationships may exist, for example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. Wherein A and B can be singular or plural. In the present invention, "at least one" refers to one or more, and "plural" refers to two or more. "At least one", "at least one of the following" or similar expressions refer to any combination of these items, including any combination of singular items or plural items. For example, "at least one of a, b, or c" or "at least one of a, b and c" can both mean: a, b, c, ab, i.e. a and b, ac, bc or abc, where a, b, c can be single or plural, respectively.

[0085] like Figure 1-Figure 3 As shown, the embodiment of the present invention provides an anonymous authentication and security aggregation method for medical data federated learning, including the following steps:

[0086] Step 1: System initialization;

[0087] Initialize system parameters and keys according to the trusted authority TA, randomly select a set of secure hash functions, and generate a master public key and a master private key;

[0088] Step 2: Federated Learning Participants U iRegistration and pseudo-identity generation;

[0089] According to the trusted institution TA, the federated learning participants U i Randomly select a temporary private key and generate a temporary public key for the pseudo-identity, which is sent to the trusted authority TA. The trusted authority TA verifies the validity of the pseudo-identity and generates complete pseudo-identity information;

[0090] Step 3: Generate key;

[0091] According to the trusted institution TA and the aggregation server Server, the pseudo identity information and the real identity information are respectively input into the trusted institution TA and the aggregation server Server to generate a long-term key;

[0092] Step 4: Model training;

[0093] The aggregation server initializes the global model to obtain global model parameters, selects users participating in federated learning and sets sequence indexes, makes the user list public and distributes the global model parameters to all hospitals’ federated learning participants U i ;

[0094] Step 5: Generate random vectors;

[0095] Generate a random vector based on its own pseudo-identity information and perform a mask operation on the locally trained global model parameters;

[0096] Step 6: Signature authentication;

[0097] The signature is authenticated based on the local gradient model, and then the signature authentication information is sent to the aggregation server Server, which verifies the validity of the signature and sends it to the federated learning participant U i Conduct certification;

[0098] Step 7: Session key generation;

[0099] Federated Learning Participants i The aggregation server generates a session key based on the pseudo-identity information for encryption and decryption of global model parameter transmission;

[0100] Step 8: The aggregation server performs security aggregation;

[0101] The aggregation server securely aggregates the global model parameter update values ​​uploaded by the hospital to obtain the federated average model parameters, and sends the federated average model parameters to the federated learning participants U of each hospital using the session key. i .

[0102] Specifically: the aggregation server performs secure aggregation and calculates the model update value after federation averaging, W i The random vectors in can cancel each other out through addition and subtraction operations, and then use the session key sk s ,i sends the updated global model W to each federated learning participant U i , return to model training for the next round of training.

[0103] like Figure 3 As shown: the system initialization specifically includes:

[0104] The trusted authority TA is given a finite cyclic additive group G on the elliptic curve with a generator of P and an order of q, and a security parameter k;

[0105] Step 11: The trusted institution TA randomly selects s∈Z q * As the master private key, the master public key is P pub =sP, where: Z q * represents a finite field, that is, the set of integers from 0 to q-1.

[0106] Step 12: The trusted institution TA selects 7 secure hash functions, where the output results of H1 and H2 are the same length as the real identity RID of the participant:

[0107] H1: G×G→{0,1} |RID|

[0108] H2: G×G×{0,1}*×G→{0,1} |RID|

[0109] H3: {0,1} * ×G→Z q *

[0110] H4: {0,1} * ×G×G→Z q *

[0111] H5: {0,1} * ×{0,1} * ×G×G×G×{0,1} * ×{0,1} * →Z q *

[0112] H6: {0,1} * ×{0,1} * ×G×G×G×{0,1} * →Z q*

[0113] H7: {0,1} * ×{0,1} * ×G→Z q *

[0114] Where: H1 represents the first hash function, H2 represents the second hash function, H3 represents the third hash function, H4 represents the fourth hash function, H5 represents the fifth hash function, H6 represents the sixth hash function, H7 represents the seventh hash function, |RID| represents the length of the real identity RID of the participant, → represents the mapping, {0,1} * Represents a string of any length composed of 0 and 1.

[0115] Step 13: The trusted institution TA publishes system parameters param = {q, P, G, P pub ,H1,H2,H3,H4,H5,H6,H7}.

[0116] like Figure 2 and Figure 3 As shown: The federated learning participants U i Registration and pseudo-identity generation specifically include:

[0117] Federated Learning Participants i Generate partial pseudo-identity and register with trusted institution TA as federated learning participant U i Generate a pseudo-identity.

[0118] like Figure 3 As shown: The federated learning participants U i Generate some pseudo identities and register them, including:

[0119] Step 21: Federated Learning Participants U i Randomly select a temporary private key x i ∈Z q * , calculate the temporary public key as X i =x i P;

[0120] Step 22: Federated Learning Participants U i Choose a random number y i ∈Z q * , calculate the temporary public key Y of the pseudo-identity i =y i P;

[0121] Step 23: Federated Learning Participants U i Calculate using real identity RID i Calculate Ai =RID i ⊕H1(y i P pub ,Y i ), and the federated learning participant U i Send(RID i ,X i ,Y i ,A i ) to the trusted institution TA;

[0122] Among them, Z q * represents a finite field, a set of integers from 0 to q-1, P represents the generator on the elliptic curve given by the trusted authority TA, H1 represents the first hash function, P pub Represents the master public key, A i Indicates RID i Signature.

[0123] like Figure 3 As shown: the trusted institution TA generates a pseudo identity for the federated learning participant Ui, specifically including:

[0124] Step 24: The trusted authority TA calculates the RID i '=A i ⊕H1(sY i ,Y i ), verify the calculated RID i 'Is it equal to the real identity RID sent? i , if they are equal, continue, real identity RID i has not been tampered with, otherwise the registration application will be rejected;

[0125] Step 25: The trusted institution TA calculates the federated learning participant U i Pseudo-identity PID i =RID i ⊕H2(sY i ,Y i ,PIV i ,P pub );

[0126] Step 26: Federated Learning Participant U i The complete pseudo identity is CPID i =(PID i ,PIV i ), where PIV i It's U i The validity period of the pseudo identity, after which the federated learning participant U i Need to re-register to generate a pseudo identity, the trusted institution TA sends CPID i To federated learning participantsi ;

[0127] Step 27: The trusted institution TA announces the federated learning participants U i List L = {X i ,PID i ,PIV i};

[0128] Among them, RIDi' represents the real identity calculated by the trusted agency TA, Ai represents the signature of RIDi, H1 represents the first hash function, H2 represents the second hash function, Yi represents the temporary public key, s represents the master private key, PIVi represents the validity period of the Ui pseudo-identity, Ppub represents the master public key, and Xi represents the temporary private key.

[0129] like Figure 3 As shown: the key generation specifically includes:

[0130] Step 31: The trusted institution TA selects a random number r i ∈Z q * , calculate U i The public key is R i =r i P, the trusted institution TA calculates B i =H3(PID i ,R i ), the trusted institution TA calculates C i =H4(PID i ,sX i ,R i ), the private key is s i =r i +B i s, sent by the trusted institution TA (s i ,R i ,B i ,C i ) to federated learning participants U i ;

[0131] Step 32: The aggregation server selects a random number x s ∈Z q * , calculate the temporary public key X s =x s P, X s Sent to the trusted institution TA, which selects a random number r s ∈Z q * , calculate the aggregation server Server public key is R s =r s P, the trusted institution TA calculates B s=H3(RID s ,R s ), the trusted institution TA calculates C s =H4(RID s ,sX s ,R s ), the private key is s s =r s +B s s, sent by the trusted institution TA (s s ,R s ,B s ,C s ) to the aggregation server Server;

[0132] Step 33: Federated Learning Participant U i Calculate C i '=H4(PID i ,x i P pub ,R i ), verify the calculated C i 'Is it equal to the sent C i , if they are equal, the key is accepted, otherwise it is rejected. The federated learning participant U i Get PID i The long-term key (s i ,R i );

[0133] Step 34: Aggregation server Server calculates C s '=H4(RID s ,x s P pub ,R s ), verify the calculated C s 'Is it equal to the sent C s If they are equal, the key is accepted, otherwise it is rejected. The aggregation server obtains the long-term key (s s ,R s );

[0134] Among them, Z q * represents a finite field, a set of integers from 0 to q-1, P represents the generator on the elliptic curve given by the trusted authority TA, H3 represents the third hash function, PID i Indicates U i The pseudo identity of H4 is the fourth hash function, s is the master private key, and X i Represents a temporary private key, P pub represents the master public key, B i Indicates U i Part of the private key, C i Indicates Ui Signature of the public key, B s Indicates a part of the private key of the aggregation server Server, RID s Indicates the true identity of the aggregation server Server, C i ’Indicates the public key signature calculated by Ui, C s Indicates the public key signature of the aggregation server Server, C s ’Indicates the public key signature calculated by the aggregation server Server.

[0135] The model training specifically includes: The aggregation server server selects users participating in federated learning for the global model (sets all model parameters to zero) and sets the sequence index id i , and publicly discloses the user list L2 = {X i , PID i , PIV i , id i} and distributes the model to all federated learning participants U i . U i Trains locally using local data and the model sent by the server to obtain w i ;

[0136] Among them, Xi represents the temporary private key, PIDi represents the pseudo-identity of Ui, and PIVi represents the validity period of the pseudo-identity of Ui.

[0137] Such as Figure 3 shown: The random vector generation specifically includes:

[0138] The federated learning participant U i Obtains the participant list L2 = {X i , PID i , PIV i , id i}, verifies whether the validity period of the user's pseudo-identity has expired, and for valid users U j Calculates the seed x of the pseudo-random number generator PRG for the federated learning participant U i = x i,j X i X j , and expands this value into a random vector: p i,j = Δ i,j ·PRG(x i,j ).

[0139] Specifically: Among them, when i > j, Δ i,j = 1; when i < j, Δ i,j = -1, U i Updates the model w i After that, for w iAdding random vectors gives: W i =w i +∑ i≠j p i,j ,

[0140] And Xi represents a temporary private key, PIDi represents the pseudo identity of Ui, PIVi represents the validity period of the pseudo identity of Ui, and idi represents a sequence index.

[0141] like Figure 3 As shown: The signature authentication specifically includes:

[0142] Step 61: Federated Learning Participant U i Sign the gradient:

[0143] Federated Learning Participants i Select a random number e i ∈Z q * , E i =e i P, federated learning participant U i Choose a random number f i ∈Z q * , F i =f i P, federated learning participant U i Calculate k i =H5(RID s ,PID i ,R i ,E i ,F i ,T i ,W i ),u i =e i +k i s i , signed by σ i =(k i ,u i ,R i ), federated learning participants U i Send(W i ,k i ,u i ,R i ,F i ,T i ,CPID i ) to the aggregation server Server;

[0144] Step 62: Aggregation Server Verification:

[0145] Aggregation Server Check PIVi Whether it is expired, get the current timestamp T now , calculate |T now -T i | value, if |T now -T i |<Δt, the signature is accepted, otherwise authentication ends. Server calculates B i '=H3(PID i ,R i ), the aggregation server calculates E i '=u i P i (R i +B i 'P pub ), the aggregation server verifies h i =H5(RID s ,PID i ,R i ,E i ',F i ,T i ,W i ) is true, if true then continue, if not true then refuse to accept the message;

[0146] Step 63: Aggregation Server Authentication:

[0147] The aggregation server selects a random number e s ∈Z q * , E s =e s P, the aggregation server Server selects a random number f s ∈Z q * , F s =f s P, the aggregation server calculates k s =H6(RID s ,PID i ,R s ,E s ,F s ,T s ),u s =e s +k s s s , the aggregation server sends (k s ,u s ,F s ,R s ,RID s ,T s) to federated learning participants U i ;

[0148] Step 64: Federated Learning Participant U i Certification:

[0149] Federated Learning Participants i Get the current timestamp T now , calculate |T now -T s | value, if |T now -T s |<Δt, then continue; otherwise, end the authentication. i Calculate B s '=H3(RID s ,R s ), federated learning participants U i Calculate E s '=u s P s (R s +B s 'P pub ), federated learning participants U i Verify i =H6(RID s ,PID i ,R s ,E s ',F s ,T s ) is established, if established, continue;

[0150] Among them, Z q * represents a finite field, a set of integers from 0 to q-1, P represents the generator on the elliptic curve given by the trusted authority TA, H5 represents the fifth hash function, PID i Indicates U i The pseudo-identity of R i Indicates U i Public key, CPID i Represents the federated learning participant U i Complete pseudo identity, H3 represents the third hash function, H6 represents the sixth hash function, k i Indicates partial signature, RID s Indicates the real identity of the aggregation server Server, T i Indicates U i The current time calculated, timestamp; Δt represents the set time difference, B i 'Server calculation part W i Signed, E. i 'Indicates the part W calculated by the Serveri Signature, h i Indicates U i Calculated part W i Signature.

[0151] like Figure 3 As shown: the session key generation specifically includes:

[0152] Step 71: Federated Learning Participant U i Session key generation:

[0153] sk i,s =H7(RID s ,PID i ,f i F s );

[0154] Step 72: Generate the aggregation server session key:

[0155] sk s,i =H7(RID s ,PID i ,f s F i );

[0156] Among them, H7 represents the seventh hash function, PID i Indicates U i The pseudo-identity, f i and f s Represents the federated learning participant U i Choose a random number, F s and F i Indicates that the aggregation server Server selects a random number f s The product of .

[0157] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined in the present invention can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown in the present invention, but should conform to the widest range consistent with the principles and novel features of the present invention.

Claims

1. An anonymous authentication and secure aggregation method for medical data federated learning, characterized in that: The steps include: Step 1: System initialization; Initialize system parameters and keys according to the trusted authority TA, randomly select a set of secure hash functions, and generate a master public key and a master private key; Step 2: Federated Learning Participants U i Registration and pseudo-identity generation; According to the trusted institution TA, the federated learning participants U i Randomly select a temporary private key and generate a temporary public key for the pseudo-identity, which is sent to the trusted authority TA. The trusted authority TA verifies the validity of the pseudo-identity and generates complete pseudo-identity information; Step 3: Generate key; According to the trusted institution TA and the aggregation server Server, the pseudo identity information and the real identity information are respectively input into the trusted institution TA and the aggregation server Server to generate a long-term key; Step 4: Model training; The aggregation server Server initializes the global model to obtain the global model parameters and distributes the global model parameters to all hospitals’ federated learning participants U i ; Step 5: Generate random vectors; Generate a random vector based on its own pseudo-identity information and perform a mask operation on the locally trained global model parameters; Step 6: Signature authentication; The signature is authenticated based on the local gradient model, and then the signature authentication information is sent to the aggregation server Server, which verifies the validity of the signature and sends it to the federated learning participant U i Conduct certification; Step 7: Session key generation; Federated Learning Participants i The aggregation server generates a session key based on the pseudo-identity information for encryption and decryption of global model parameter transmission; Step 8: The aggregation server performs security aggregation; The aggregation server securely aggregates the global model parameter update values ​​uploaded by the hospital to obtain the federated average model parameters, and sends the federated average model parameters to the federated learning participants U of each hospital using the session key. i .

2. According to claim 1, the anonymous authentication and secure aggregation method for medical data federated learning is characterized in that: The system initialization specifically includes: The trusted authority TA is given a finite cyclic additive group G on the elliptic curve with a generator of P and an order of q, and a security parameter k; Step 11: The trusted institution TA randomly selects s∈Z q * As the master private key, the master public key is P pub =sP, where: Z q * represents a finite field, that is, the set of integers from 0 to q-1. Step 12: The trusted institution TA selects 7 secure hash functions, where the output results of H1 and H2 are the same length as the real identity RID of the participant: H1: G×G→{0,1} |RID| H2:G×G×{0,1} * ×G→{0,1} |RID| <h2 style=";text-align:left;direction:ltr">H3: {0,1}<h2 style=";text-align:left;direction:ltr"> * <h2 style=";text-align:left;direction:ltr"> ×G→Z<h2 style=";text-align:left;direction:ltr"> q <h2 style=";text-align:left;direction:ltr"> * H4: {0,1} * ×G×G→Z q * H5:{0,1} * ×{0,1} * ×G×G×G×{0,1} * ×{0,1} * →Z q * H6:{0,1} * ×{0,1} * ×G×G×G×{0,1} * →Z q * H7:{0,1} * ×{0,1} * ×G→Z q * Where: H1 represents the first hash function, H2 represents the second hash function, H3 represents the third hash function, H4 represents the fourth hash function, H5 represents the fifth hash function, H6 represents the sixth hash function, H7 represents the seventh hash function, |RID| represents the length of the real identity RID of the participant, → represents the mapping, {0,1} * Represents a string of any length composed of 0 and 1. Step 13: The trusted institution TA publishes system parameters param = {q, P, G, P pub ,H1,H2,H3 H4,H5,H6,H7}.

3. According to claim 1, the anonymous authentication and secure aggregation method for medical data federated learning is characterized in that: The federated learning participants U i Registration and pseudo-identity generation specifically include: Federated Learning Participants i Generate partial pseudo-identity and register with trusted institution TA as federated learning participant U i Generate a pseudo-identity.

4. According to claim 3, the anonymous authentication and secure aggregation method for medical data federated learning is characterized in that: The federated learning participants U i Generate some pseudo identities and register them, including: Step 21: Federated Learning Participants U i Randomly select a temporary private key x i ∈Z q * , calculate the temporary public key as X i =x i P; Step 22: Federated Learning Participants U i Choose a random number y i ∈Z q * , calculate the temporary public key Y of the pseudo-identity i =y i P; Step 23: Federated Learning Participant U i Calculate using real identity RID i Calculate A i =RID i ⊕H1(y i P pub , Y i ), and the federated learning participant U i Send(RID i ,X i ,Y i ,A i ) to the trusted institution TA; Among them, Z q * represents a finite field, a set of integers from 0 to q-1, P represents the generator on the elliptic curve given by the trusted authority TA, H1 represents the first hash function, P pub Represents the master public key, A i Indicates RID i Signature.

5. According to claim 3, the anonymous authentication and secure aggregation method for medical data federated learning is characterized in that: The trusted institution TA is a federated learning participant U i Generating a pseudo identity specifically includes: Step 24: The trusted authority TA calculates the RID i '=A i ⊕H1(sY i ,Y i ), verify the calculated RID i 'Is it equal to the real identity RID sent? i , if they are equal, continue, real identity RID i has not been tampered with, otherwise the registration application will be rejected; Step 25: The trusted institution TA calculates the federated learning participant U i Pseudo-identity PID i =RID i ⊕H2(sY i ,Y i ,PIV i ,P pub ); Step 26: Federated Learning Participants U i The complete pseudo identity is CPID i =(PID i ,PIV i ), PIV i It's U i The validity period of the pseudo identity, after which the federated learning participant U i Need to re-register to generate a pseudo identity, the trusted institution TA sends CPID i To federated learning participants i ; Step 27: The trusted institution TA announces the federated learning participants U i List L = {X i ,PID i ,PIV i }; Among them, RID i ' represents the real identity calculated by the trusted institution TA, A i Indicates RID i The signature of H1 represents the first hash function, H2 represents the second hash function, and Y i represents a temporary public key, s represents the master private key, PIV i Indicates U i The validity period of the pseudo-identity, P pub Represents the master public key, X i Represents a temporary private key.

6. The anonymous authentication and secure aggregation method for medical data federated learning according to claim 1, characterized in that: The key generation specifically includes: Step 31: The trusted institution TA selects a random number r i ∈Z q * , calculate U i The public key is R i =r i P, the trusted institution TA calculates B i =H3(PID i ,R i ), the trusted institution TA calculates C i =H4(PID i ,sX i ,R i ), the private key is s i =r i +B i s, sent by the trusted institution TA (s i ,R i ,B i ,C i ) to federated learning participants U i ; Step 32: The aggregation server selects a random number x s ∈Z q * , calculate the temporary public key X s =x s P, X s Sent to the trusted institution TA, which selects a random number r s ∈Z q * , calculate the aggregation server Server public key is R s =r s P, the trusted institution TA calculates B s =H3(RID s ,R s ), the trusted institution TA calculates C s =H4(RID s ,sX s ,R s ), the private key is s s =r s +B s s, sent by the trusted institution TA (s s ,R s ,B s ,C s ) to the aggregation server Server; Step 33: Federated Learning Participant U i Calculate C i '=H4(PID i ,x i P pub ,R i ), verify the calculated C i 'Is it equal to the sent C i , if they are equal, the key is accepted, otherwise it is rejected. The federated learning participant U i Get PID i The long-term key (s i ,R i ); Step 34: Aggregation server Server calculates C s '=H4(RID s ,x s P pub ,R s ), verify the calculated C s 'Is it equal to the sent C s If they are equal, the key is accepted, otherwise it is rejected. The aggregation server obtains the long-term key (s s ,R s ). Among them, Z q * represents a finite field, a set of integers from 0 to q-1, P represents the generator on the elliptic curve given by the trusted authority TA, H3 represents the third hash function, PID i Indicates U i The pseudo identity of H4 is the fourth hash function, s is the master private key, and X i Represents a temporary private key, P pub represents the master public key, B i Indicates U i Part of the private key, C i Indicates U i The signature of the public key, B s Represents part of the aggregation server's Server private key, RID s Indicates the real identity of the aggregation server Server, C i ' represents the public key signature calculated by Ui, C s Indicates the public key signature of the aggregation server Server, C s 'Indicates the public key signature calculated by the aggregation server Server.

7. The anonymous authentication and secure aggregation method for medical data federated learning according to claim 1, characterized in that: The model training specifically includes: The aggregation server selects the users who participate in federated learning with the global model (sets all model parameters to zero) and sets the sequence index idi, publishes the user list L2 = {Xi, PIDi, PIVi, idi} and distributes the model to all federated learning participants U i ; U i Use local data and the model sent by the server to train locally to get w i ; Among them, X i Indicates a temporary private key, PID i Indicates U i Pseudo-identity, PIV i Indicates U i The validity period of the pseudo-identity.

8. The anonymous authentication and secure aggregation method for medical data federated learning according to claim 1, characterized in that: The random vector generation specifically includes: Federated Learning Participant U i Obtain the participant list L2 = {X i , PID i , PIV i , id i}, verify whether the validity period of the user's pseudo-identity has expired. For the valid user U j Calculate the seed x of the pseudo-random number generator PRG for federated learning participant U i i,j = x i X j , and expand this value into a random vector: p i,j = Δ i,j ·PRG(x i,j ), where, when i > j, Δ i,j = 1; when i < j, Δ i,j = -1, U i After updating the model W i , add the random vector to W i to get: W i = w i + ∑ i≠j p i,j , and X i represents the temporary private key, PID i represents the pseudo-identity of U i , PIV i represents the validity period of the pseudo-identity of U i , and id i represents the sequence index.​ 9. The anonymous authentication and secure aggregation method for medical data federated learning according to claim 1, characterized in that: The signature authentication specifically includes: Step 61: Federated Learning Participant U i Sign the gradient: Federated Learning Participants i Choose a random number e i ∈Z q * , E i =e i P, federated learning participant U i Choose a random number f i ∈Z q * , F i =f i P, federated learning participant U i Calculate k i =H5(RID s ,PID i ,R i ,E i ,F i ,T i ,W i ),u i =e i +k i s i , signed by σ i =(k i ,u i ,R i ), federated learning participants U i Send(W i ,k i ,u i ,R i ,F i ,T i ,CPID i ) to the aggregation server Server; Step 62: Aggregation Server Verification: Aggregation Server Check PIV i Whether it is expired, get the current timestamp T now , calculate |T now -T i | value, if |T now -T i |<Δt, the signature is accepted; otherwise, authentication ends and the Server calculates B i '=H3(RID i ,R i ), the aggregation server calculates E i '=u i P i (R i +B i 'P pub ), the aggregation server verifies h i =H5(RID s ,PID i ,R i ,E i ',F i ,T i ,W i ) is true, if true, continue; if not, refuse to accept the message; Step 63: Aggregation server authentication: The aggregation server Server selects a random number e s ∈Z q * , E s =e s P, the aggregation server Server selects a random number f s ∈Z q * , F s =f s P, the aggregation server calculates k s =H6(RID s ,PID i ,R s ,E s ,F s ,T s ),u s =e s +k s s s , the aggregation server sends (k s ,u s ,F s ,R s ,RID s ,T s ) to federated learning participants U i ; Step 64: Federated Learning Participant U i Certification: Federated Learning Participants i Get the current timestamp T now , calculate |T now -T s | value, if |T now -T s |<Δt, then continue; otherwise, end the authentication. i Calculate B s '=H3(RID s ,R s ), federated learning participants U i Calculate E s '=u s P s (R s +B s 'P pub ), federated learning participants U i Verify i =H6(RID s , PID i ,R s ,E s ',F s ,T s ) is established, if established, continue; Among them, Z q * represents a finite field, a set of integers from 0 to q-1, P represents the generator on the elliptic curve given by the trusted authority TA, H5 represents the fifth hash function, PID i Indicates U i The pseudo-identity of R i Indicates U i Public key, CPID i Represents the federated learning participant U i Complete pseudo identity, H3 represents the third hash function, H6 represents the sixth hash function, k i Indicates partial signature, RID s Indicates the real identity of the aggregation server Server, T i Indicates U i The current time calculated, timestamp; Δt represents the set time difference, B i 'Server calculation part W i Signed, E. i 'Indicates the part W calculated by the Server i Signature, h i Indicates U i Calculated part W i Signature.

10. The anonymous authentication and secure aggregation method for medical data federated learning according to claim 1, characterized in that: The session key generation specifically includes: Step 71: Federated Learning Participant U i Session key generation: sk i,s =H7(RID s ,PID i ,f i F s ); Step 72: Generate the aggregation server session key: sk s,i =H7(RID s ,PID i ,f s F i ); Among them, H7 represents the seventh hash function, PID i Indicates U i The pseudo-identity, f i and f s Represents the federated learning participant U i Choose a random number, F s and F i Indicates that the aggregation server Server selects a random number f s The product of .

Citation Information

Patent Citations

  • Anonymous authentication method and system

    CN112364331A

  • Federal learning-based data transmission method and device, equipment and medium

    CN114186213A

  • Asynchronous federated learning privacy protection method and system, medium, equipment and terminal

    CN115277015A

  • Local model parameter aggregation method for federal learning

    CN116702191A

  • Communication-efficient federated learning system with data protection and data security

    DE202023107184U1

Cited By

  • Federal learning model security aggregation method based on differential privacy

    CN122226446A