A data encryption method based on privacy computing

Abstract

The present invention relates to the technical field of data encryption, and discloses a data encryption method based on privacy computing, including: obtaining user submitted data, encoding the user submitted data using a word2vec model; encrypting the encoding vector using a homomorphic encryption algorithm, and transmitting the encryption result to a remote server; when performing a query calculation operation, converting the query field into an encoding vector using a word2vec model, and transmitting the same to the remote server after homomorphic encryption, and the remote server performs homomorphic calculation of the average Euclidean distance between the query field encoding result and the stored encoding vector ciphertext; if the calculated average Euclidean distance is less than a specified threshold, the remote server returns the encoding vector ciphertext corresponding to the user encrypted data, and the user server decrypts the encrypted data and performs an inverse operation using the word2vec model to obtain the original data. The method of the present invention avoids the problem of data leakage by performing calculation processing on the encrypted data.

Description

Technical Field

[0001] The present invention relates to the technical field of data encryption, and in particular to a data encryption method based on privacy computing. Background Art

[0002] The data submitted by users to banks is sensitive. In the process of storing the submitted data on a remote server through traditional encryption, it is vulnerable to man-in-the-middle attacks and data leakage, which has a great impact on the user's personal privacy and the bank's reputation.

[0003] In view of this, the present invention proposes a data encryption method based on privacy computing, which avoids data leakage problems by performing calculations on encrypted data. Summary of the invention

[0004] The present invention provides a data encryption method based on privacy computing, the purpose of which is to (1) realize the encoding and encryption processing of private data; and (2) realize the homomorphic calculation of encrypted ciphertext, thereby realizing the query of private data.

[0005] To achieve the above object, the present invention provides a data encryption method based on privacy computing, comprising the following steps:

[0006] The user submits the user information data field on the user server; a word2vec model is constructed and deployed on the user server, and the user server encodes the received user information data field using the word2vec model to obtain encoding vectors of different user information data fields;

[0007] The user server encrypts the encoding vector using a homomorphic encryption algorithm and transmits the encryption result to a remote server, which stores the ciphertext of the encoding vector.

[0008] When performing a query calculation operation, the user enters a query field on the user server, and the user server uses the word2vec model to convert the query field into a query field encoding vector, and homomorphically encrypts the query field encoding vector and transmits it to the remote server. The remote server receives the ciphertext of the query field and homomorphically calculates the average Euclidean distance between the query field ciphertext and the stored encoding vector ciphertext;

[0009] According to the homomorphic calculation results, if the minimum distance between the query field ciphertext and the stored encoding vector ciphertext is less than the specified threshold, it means that the remote server has stored the query content, and the remote server returns the encoding vector ciphertext that meets the specified threshold to the user server. The user server decrypts the received encoding vector ciphertext and uses the word2vec model to perform an inverse operation on the decryption result to obtain the query data and return the query data; if the minimum distance between the query field ciphertext and the stored encoding vector ciphertext is greater than the specified threshold, it means that the remote server does not store the query content, and the user server directly returns no query result.

[0010] As a further improvement method of the present invention:

[0011] After the user submits the user information data field at the user service end, the method further includes:

[0012] Construct a user feature dictionary and perform one-hot encoding on the user information data field based on the user feature dictionary. The encoding result is x={(x 1,1 ,x 1,2 ,…,x 1,L ),(x 2,1 ,x 2,2 ,…,x 2,L ),…,(x n,1 ,x n,2 ,…,x n,L )}, where L represents the size of the user feature dictionary, n represents the number of categories of the user information data field, (x n,1 ,x n,2 ,…,x n,L ) represents the one-hot encoding result of the n-th user information data field, in which only one unit x n,i The value of is 1, indicating the position of the user information data field of this category in the dictionary, and the values ​​of other units are all 0. In a specific embodiment of the present invention, the user information data field category includes user account, name, home address, educational background and income, etc.

[0013] The word2vec model is constructed, including:

[0014] The word2vec model includes an output layer, a hidden layer, and an output layer. The one-hot encoding result x of the user information data field is used as the input value of the word2vec model input layer. The calculation result of the hidden layer of the word2vec model is:

[0015] h i =w h T x i

[0016] in:

[0017] x i represents the one-hot encoding result of the i-th user information data field, h i Indicates the corresponding hidden layer calculation results;

[0018] w h represents the weight matrix in the hidden layer, w h The size of is L×H, where L represents the size of the user feature dictionary and H represents the size of the hidden layer;

[0019] The calculation results of the hidden layer are input to the output layer. For the i-th type of user information data field, the output result of the output layer is:

[0020]

[0021] in:

[0022] w y represents the weight matrix in the output layer, w y The size is H×L;

[0023] The resulting encoding vector is (y1,y2,…,y i ,…,y n ).

[0024] The word2vec model constructed by the training includes:

[0025] According to the constructed user feature dictionary, a training corpus containing feature words in the user feature dictionary is constructed;

[0026] Determine the objective function for training the word2vec model:

[0027]

[0028] in:

[0029] M is the number of characteristic words in the training corpus, x ′ m Represents the characteristic words in the training corpus;

[0030] θ(w h T ,w y T ) are the parameters of the word2vec model, including the weight matrix w of the hidden layer h T And the weight matrix w of the output layer y T ;

[0031] Determine the update formula for word2vec model parameters:

[0032]

[0033] in:

[0034] α represents the learning rate, which is set to 0.1;

[0035] Iterate the word2vec model parameters until the objective function value reaches the maximum, and use the updated parameters corresponding to the maximum objective function value as the training parameters of the word2vec model.

[0036] The user server encrypts the encoding vector using a homomorphic encryption algorithm, including:

[0037] The encoding vector is encrypted using a homomorphic encryption algorithm, and the homomorphic encryption algorithm process is as follows:

[0038] For the encoding vector (y1,y2,…,y i ,…,y n ) in any word vector y i , using the public key PK for homomorphic encryption:

[0039] c i =PK(y i )

[0040] in:

[0041] PK(·) means using the public key PK of the user server to encrypt the encoded data into ciphertext;

[0042] Get the encrypted ciphertext (c1, c2, ..., c i ,…,c n ).

[0043] The homomorphic calculation of the average Euclidean distance between the query field ciphertext and the stored encoding vector ciphertext includes:

[0044] The query field ciphertext c * and the ciphertext of the encoding vector (r1, r2, …, r j ,…,r q ) performs homomorphic calculation of the average Euclidean distance, where r j =(c j1 ,c j2 ,…,c ji ,…,c jn ) represents the jth coded vector ciphertext, q represents the number of coded vector ciphertexts stored in the remote server, where the homomorphic calculation process of the average Euclidean distance is:

[0045] 1) The remote server performs the following operations to obtain the ciphertext Euclidean distance calculation result:

[0046]

[0047] in:

[0048] y * represents the encoding vector of the query field, e represents the number of words in the query field, represents the encoding vector of the e-th word in the query field;

[0049] y je represents the encoding vector of the e-th word corresponding to the j-th encoding vector ciphertext;

[0050] 2) Using the rotation method to ds i Sum the elements in :

[0051]

[0052] in:

[0053] rotate(ds q ,2 k ) means 2 k The result of calculating the ciphertext Euclidean distance of the initial element ds q ;

[0054]

[0055] in:

[0056] d q represents the Euclidean distance between the query field encoding vector and the encoding vector corresponding to the qth encoding vector ciphertext, PK(0,0,…,d q ) represents the Euclidean distance between the ciphertext of the query field encoding vector and the ciphertext of the qth encoding vector;

[0057] 3) Combine and sum the Euclidean distance:

[0058] cd * =PK(d1,d2,…,d q )

[0059] in:

[0060] cd * The element value in represents the distance between the ciphertext of the encoding vector of the query field and the ciphertext of the encoding vector in the remote server.

[0061] The private data query result is returned according to the homomorphic computation result, including:

[0062] Traversing CD * The Euclidean distance value d ini , find the minimum value d min , if the minimum distance PK(d min ) is less than the specified threshold It means that the remote server stores the query content. The remote server returns the coded vector ciphertext that meets the specified threshold to the user server. The user server decrypts the received coded vector ciphertext and uses the word2vec model to perform the inverse operation on the decryption result to obtain the query data and return the query data. If the minimum distance PK (d min ) is greater than the specified threshold This means that the remote server does not store the query content, and the user server directly returns no query results.

[0063] Compared with the prior art, the present invention proposes a data encryption method based on privacy computing, which has the following advantages:

[0064] First, this scheme proposes an encoding and encryption process for private data. By using the one-hot encoding result x of the user information data field as the input value of the input layer of the word2vec model, the calculation result of the hidden layer of the word2vec model is:

[0065] h i =w h T x i

[0066] Where: x i represents the one-hot encoding result of the i-th user information data field, h i Indicates the corresponding hidden layer calculation result; w h represents the weight matrix in the hidden layer, w h The size of is L×H, L represents the size of the user feature dictionary, and H represents the size of the hidden layer; the calculation result of the hidden layer is input to the output layer. For the i-th user information data field, the output result of the output layer is:

[0067]

[0068] Where: w y represents the weight matrix in the output layer, w y The size of is H×L; the resulting encoding vector is (y1,y2,…,y i ,…,y n ). At the same time, this scheme uses a homomorphic encryption algorithm to encrypt the encoding vector. The homomorphic encryption algorithm process is as follows: for the encoding vector (y1, y2, ..., y i ,…,yn ) in any word vector y i , using the public key PK for homomorphic encryption:

[0069] c i =PK(y i )

[0070] Where: PK(·) means using the public key PK of the user server to encrypt the encoded data into ciphertext; the encrypted ciphertext (c1, c2, …, c i ,…,c n ). Compared with the traditional solution, this solution first uses a pre-built user feature dictionary to one-hot encode the user's privacy data. If the constructed user feature dictionary is not known, the encoding result cannot be decoded, and the word2vec model is used to encode the word vector. The encoding result is encrypted using the homomorphic encryption algorithm to obtain the encrypted ciphertext, thereby using multiple privacy data protection methods to achieve user privacy data encryption.

[0071] At the same time, this scheme proposes a privacy data query scheme by querying the ciphertext c of the field * and the ciphertext of the encoding vector (r1, r2, …, r j ,…,r q ) performs homomorphic calculation of the average Euclidean distance, where r j =(c j1 ,c j2 ,…,c ji ,…,c jn ) represents the jth coded vector ciphertext, q represents the number of coded vector ciphertexts stored in the remote server, wherein the homomorphic calculation process of the average Euclidean distance is: the remote server performs the following operation to obtain the ciphertext Euclidean distance calculation result:

[0072]

[0073] Where: y * represents the encoding vector of the query field, e represents the number of words in the query field, represents the encoding vector of the e-th word in the query field; y je represents the encoding vector of the e-th word corresponding to the j-th encoding vector ciphertext; ds is rotated by i Sum the elements in :

[0074]

[0075] Where: rotate(ds q ,2 k ) means 2 k The result of calculating the ciphertext Euclidean distance of the initial element dsq ;

[0076]

[0077] Where: d q represents the Euclidean distance between the query field encoding vector and the encoding vector corresponding to the qth encoding vector ciphertext, PK(0,0,…,d q ) represents the Euclidean distance between the ciphertext of the query field encoding vector and the ciphertext of the qth encoding vector; the Euclidean distance obtained by merging and summing is:

[0078] cd * =PK(d1,d2,…,d q )

[0079] Where: cd * The element value in represents the distance between the ciphertext of the encoding vector of the query field and the ciphertext of the encoding vector in the remote server. * The Euclidean distance value d in i , find the minimum value d min , if the minimum distance PK(d min ) is less than the specified threshold It means that the remote server stores the query content. The remote server returns the coded vector ciphertext that meets the specified threshold to the user server. The user server decrypts the received coded vector ciphertext and uses the word2vec model to perform the inverse operation on the decryption result to obtain the query data and return the query data. If the minimum distance PK (d min ) is greater than the specified threshold This means that the remote server does not store the query content, and the user server directly returns no query results, thereby realizing user privacy data query. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] Figure 1 A flowchart of a data encryption method based on privacy computing provided by one embodiment of the present invention;

[0081] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0082] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0083] S1: Obtain user submitted data, encode the user submitted data using the word2vec model, and obtain the encoding vector of the user submitted data.

[0084] After the user submits the user information data field at the user server in step S1, the method further includes:

[0085] Construct a user feature dictionary and perform one-hot encoding on the user information data field based on the user feature dictionary. The encoding result is x={(x 1,1 ,x 1,2 ,…,x 1,L ),(x 2,1 ,x 2,2 ,…,x 2,L ),…,(x n,1 ,x n,2 ,…,x n,L )}, where L represents the size of the user feature dictionary, n represents the number of categories of the user information data field, (x n,1 ,x n,2 ,…,x n,L ) represents the one-hot encoding result of the n-th user information data field, in which only one unit x n,i The value of is 1, indicating the position of the user information data field of this category in the dictionary, and the values ​​of other units are all 0. In a specific embodiment of the present invention, the user information data field category includes user account, name, home address, educational background and income, etc.

[0086] The word2vec model is constructed in the S1 step, including:

[0087] The word2vec model includes an output layer, a hidden layer, and an output layer. The one-hot encoding result x of the user information data field is used as the input value of the word2vec model input layer. The calculation result of the hidden layer of the word2vec model is:

[0088] h i =w h T x i

[0089] in:

[0090] x i represents the one-hot encoding result of the i-th user information data field, h i Indicates the corresponding hidden layer calculation results;

[0091] w h represents the weight matrix in the hidden layer, w h The size of is L×H, where L represents the size of the user feature dictionary and H represents the size of the hidden layer;

[0092] The calculation results of the hidden layer are input to the output layer. For the i-th type of user information data field, the output result of the output layer is:

[0093]

[0094] in:

[0095] w y represents the weight matrix in the output layer, w y The size is H×L;

[0096] The resulting encoding vector is (y1,y2,…,y i ,…,y n ).

[0097] The word2vec model constructed by training in step S1 includes:

[0098] According to the constructed user feature dictionary, a training corpus containing feature words in the user feature dictionary is constructed;

[0099] Determine the objective function for training the word2vec model:

[0100]

[0101] in:

[0102] M is the number of characteristic words in the training corpus, x ′ m Represents the characteristic words in the training corpus;

[0103] θ(w h T ,w y T ) are the parameters of the word2vec model, including the weight matrix w of the hidden layer h T And the weight matrix w of the output layer y T ;

[0104] Determine the update formula for word2vec model parameters:

[0105]

[0106] in:

[0107] α represents the learning rate, which is set to 0.1;

[0108] Iterate the word2vec model parameters until the objective function value reaches the maximum, and use the updated parameters corresponding to the maximum objective function value as the training parameters of the word2vec model.

[0109] S2: Encrypt the encoding vector using a homomorphic encryption algorithm and transmit the encryption result to a remote server, which stores the encoding vector ciphertext of the user data.

[0110] In step S2, the user server encrypts the encoding vector using a homomorphic encryption algorithm, including:

[0111] The encoding vector is encrypted using a homomorphic encryption algorithm, and the homomorphic encryption algorithm process is as follows:

[0112] For the encoding vector (y1,y2,…,y i ,…,y n ) in any word vector y i , using the public key PK for homomorphic encryption:

[0113] c i =PK(y i )

[0114] in:

[0115] PK(·) means using the public key PK of the user server to encrypt the encoded data into ciphertext;

[0116] Get the encrypted ciphertext (c1, c2, ..., c i ,…,c n ), and transmits the encryption result to a remote server, which stores the encoding vector ciphertext of the user data.

[0117] S3: When performing query calculation operations, the word2vec model is used to convert the query field into an encoding vector, which is homomorphically encrypted and transmitted to the remote server. The remote server receives the encoding result of the query field and performs homomorphic calculation of the average Euclidean distance between the encoding result of the query field and the stored encoding vector ciphertext.

[0118] In the step S3, the query field ciphertext and the stored encoding vector ciphertext are homomorphically calculated based on the average Euclidean distance, including:

[0119] The query field ciphertext c * and the ciphertext of the encoding vector (r1, r2, …, r j ,…,r q ) performs homomorphic calculation of the average Euclidean distance, where r j =(c j1 ,c j2 ,…,c ji ,…,c jn ) represents the jth coded vector ciphertext, q represents the number of coded vector ciphertexts stored in the remote server, where the homomorphic calculation process of the average Euclidean distance is:

[0120] 1) The remote server performs the following operations to obtain the ciphertext Euclidean distance calculation result:

[0121]

[0122] in:

[0123] y * represents the encoding vector of the query field, e represents the number of words in the query field, represents the encoding vector of the e-th word in the query field;

[0124] y je represents the encoding vector of the e-th word corresponding to the j-th encoding vector ciphertext;

[0125] 2) Using the rotation method to ds i Sum the elements in :

[0126]

[0127] in:

[0128] rotate(ds q ,2 k ) means 2 k The result of calculating the ciphertext Euclidean distance of the initial element ds q ;

[0129]

[0130] in:

[0131] d q represents the Euclidean distance between the query field encoding vector and the encoding vector corresponding to the qth encoding vector ciphertext, PK(0,0,…,d q ) represents the Euclidean distance between the ciphertext of the query field encoding vector and the ciphertext of the qth encoding vector;

[0132] 3) Combine and sum the Euclidean distance:

[0133] cd * =PK(d1,d2,…,d q )

[0134] in:

[0135] cd * The element value in represents the distance between the ciphertext of the encoding vector of the query field and the ciphertext of the encoding vector in the remote server.

[0136] S4: If the calculated average Euclidean distance is less than the specified threshold, the remote server returns the coded vector ciphertext corresponding to the user's encrypted data. The user server decrypts the encrypted data and performs an inverse operation using the word2vec model to obtain the original data.

[0137] In step S4, based on the homomorphic computation result, the private data query result is returned, including:

[0138] Traversing CD * The Euclidean distance value d in i , find the minimum value d min , if the minimum distance PK(d min ) is less than the specified threshold It means that the remote server stores the query content. The remote server returns the coded vector ciphertext that meets the specified threshold to the user server. The user server decrypts the received coded vector ciphertext and uses the word2vec model to perform the inverse operation on the decryption result to obtain the query data and return the query data. If the minimum distance PK (d min ) is greater than the specified threshold This means that the remote server does not store the query content, and the user server directly returns no query results.

[0139] It should be noted that the serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages and disadvantages of the embodiments. And the terms "including", "comprising" or any other variants thereof in this article are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "including a ..." does not exclude the presence of other identical elements in the process, device, article or method including the element.

[0140] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0141] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A data encryption method based on privacy computing, characterized in that: The method comprises: S11: The user submits the user information data field on the user server; a word2vec model is constructed and deployed on the user server, and the user server encodes the received user information data field using the word2vec model to obtain encoding vectors of different user information data fields; Build a user feature dictionary and perform one-hot encoding on the user information data field based on the user feature dictionary. The encoding result is , where L represents the size of the user feature dictionary, n represents the number of categories of the user information data field, Represents the one-hot encoding result of the n-th category user information data field, in which only one unit The value of is 1, indicating the position of the user information data field in the dictionary, and the values ​​of other cells are 0; S12: The user server encrypts the encoding vector using a homomorphic encryption algorithm and transmits the encryption result to the remote server, which stores the ciphertext of the encoding vector; The homomorphic encryption algorithm process is as follows: For the encoding vector Any word vector in , using the public key PK for homomorphic encryption: ； in: It means using the public key PK of the user server to encrypt the encoded data into ciphertext; Get encrypted ciphertext ; S13: When performing a query calculation operation, the user inputs a query field on the user server, and the user server uses the word2vec model to convert the query field into a query field encoding vector, and homomorphically encrypts the query field encoding vector and transmits it to the remote server. The remote server receives the ciphertext of the query field, and homomorphically calculates the average Euclidean distance between the query field ciphertext and the stored encoding vector ciphertext; S14: According to the homomorphic calculation result, if the minimum distance between the query field ciphertext and the stored encoding vector ciphertext is less than the specified threshold, it means that the remote server has stored the query content, and the remote server returns the encoding vector ciphertext that meets the specified threshold to the user server. The user server decrypts the received encoding vector ciphertext, and uses the word2vec model to perform an inverse operation on the decryption result to obtain the query data, and returns the query data; if the minimum distance between the query field ciphertext and the stored encoding vector ciphertext is greater than the specified threshold, it means that the remote server does not store the query content, and the user server directly returns no query result.

2. A data encryption method based on privacy computing as claimed in claim 1, characterized in that: The word2vec model is constructed, including: The word2vec model includes an output layer, a hidden layer, and an output layer. The one-hot encoding result x of the user information data field is used as the input value of the word2vec model input layer. The calculation result of the hidden layer of the word2vec model is: ； in: represents the one-hot encoding result of the i-th user information data field, Indicates the corresponding hidden layer calculation results; represents the weight matrix in the hidden layer, The size is , L represents the size of the user feature dictionary, and H represents the size of the hidden layer; The calculation results of the hidden layer are input to the output layer. For the i-th type of user information data field, the output result of the output layer is: ； in: represents the weight matrix in the output layer, The size is ; The resulting encoding vector is .

3. A data encryption method based on privacy computing as described in claim 2, characterized in that: The word2vec model constructed by training includes: According to the constructed user feature dictionary, a training corpus containing feature words in the user feature dictionary is constructed; Determine the objective function for training the word2vec model: ； in: M is the number of characteristic words in the training corpus, Represents the characteristic words in the training corpus; The parameters of the word2vec model, including the weight matrix of the hidden layer And the weight matrix of the output layer ; Determine the update formula for word2vec model parameters: ； in: Represents the learning rate, set it to 0.1; Iterate the word2vec model parameters until the objective function value reaches the maximum, and use the updated parameters corresponding to the maximum objective function value as the training parameters of the word2vec model.

4. A data encryption method based on privacy computing as claimed in claim 1, characterized in that: The homomorphic calculation of the average Euclidean distance between the query field ciphertext and the stored encoding vector ciphertext includes: Encrypt the query field The encrypted vector stored in the remote server Perform homomorphic calculation of the average Euclidean distance, where represents the jth coded vector ciphertext, q represents the number of coded vector ciphertexts stored in the remote server, and the homomorphic calculation process of the average Euclidean distance is: 1) The remote server performs the following operations to obtain the ciphertext Euclidean distance calculation result: ； ； in: represents the encoding vector of the query field, e represents the number of words in the query field, represents the encoding vector of the e-th word in the query field; represents the code vector of the e-th word corresponding to the j-th code vector ciphertext; 2) Using the rotation method Sum the elements in : ； ； in: Indicates The result of calculating the ciphertext Euclidean distance of the initial element ; ； in: represents the Euclidean distance between the query field encoding vector and the encoding vector corresponding to the qth encoding vector ciphertext, represents the Euclidean distance between the ciphertext of the query field encoding vector and the ciphertext of the qth encoding vector; 3) Euclidean distance obtained by merging and summing: ； in: The element value in represents the distance between the ciphertext of the encoding vector of the query field and the ciphertext of the encoding vector in the remote server.

5. A data encryption method based on privacy computing as claimed in claim 4, characterized in that: The private data query result is returned according to the homomorphic computation result, including: Traversal The Euclidean distance value in , find the minimum value among them , if the minimum distance between the query field ciphertext and the stored encoding vector ciphertext Less than the specified threshold , it means that the remote server stores the query content. The remote server returns the coded vector ciphertext that meets the specified threshold to the user server. The user server decrypts the received coded vector ciphertext and uses the word2vec model to perform the inverse operation on the decryption result to obtain the query data and return the query data. If the minimum distance between the query field ciphertext and the stored coded vector ciphertext is Greater than a specified threshold , it means that the remote server does not store the query content, and the user server directly returns no query results.

Images