Business collaborative processing method, device, medium and equipment based on privacy protection
Patent Information
- Application Number
- CN202111580145.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-12-22
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2041-12-22
AI Technical Summary
In business collaboration processing, how to achieve data privacy protection without leaking the business data held by various business parties, especially in business processing such as feature data quality analysis and AI model training, it is difficult to effectively complete the existing technology. The collaborative sorting of business data leads to limited business processing.
By building an initial auxiliary sequence whose length is the power of 2 of the merged sequence length, element filling and iterative splitting and interchange processing are performed, the sorting of business data is gradually realized, data privacy is ensured, and plaintext sorting is performed locally to reduce the number of Communication requirements.
It realizes efficient completion of business collaborative processing without leaking data privacy, reduces the complexity of communication and data processing, simplifies the business collaborative processing process, is suitable for parallel processing, and significantly improves processing efficiency.
Smart Images

Figure CN114254355B_ABST
Abstract
Description
Technical Field
[0001] The present specification is applied to the field of business collaborative processing technology, and specifically relates to a business collaborative processing method, device, medium and equipment based on privacy protection. Background Art
[0002] Secure multi-party computation (MPC) mainly solves the problem of how two or more parties can successfully complete the calculation without revealing their inputs to each other. The current application prospects of MPC are becoming more and more extensive, and the demand for business collaboration between enterprises, units and individuals is becoming increasingly strong. A typical scenario is that in the rapid development of artificial intelligence, the demand for data privacy has become increasingly strong. In business processing such as feature data quality analysis, AI model training, and AI model application, the required business data cannot be obtained due to privacy compliance reasons, resulting in the inability to complete the corresponding business processing. At present, attempts are being made to use MPC to solve the problem of privacy protection in business processing, that is, how to have multiple business parties collaborate to perform business processing without directly exposing plaintext data. Summary of the invention
[0003] The purpose of the embodiments of this specification is to provide a business collaborative processing method, device, medium and equipment based on privacy protection, which can more easily and efficiently complete the collaborative sorting of business information involved in the business collaborative processing process without exposing the business information held by each business party, thereby more easily and efficiently realizing business collaborative processing based on privacy protection.
[0004] This specification provides a business collaborative processing method, device, medium and equipment based on privacy protection, which are implemented in the following ways:
[0005] A business collaborative processing method based on privacy protection is applied to any business party of two business parties holding sequences to be sorted in business collaborative processing; the method comprises: cooperating with the other business party of the two business parties to construct an initial auxiliary sequence based on the length of the merged sequence; wherein the sequence length of the initial auxiliary sequence is a value greater than or equal to the length of the merged sequence and is a power of 2; the merged sequence length is the sum of the lengths of the first privacy sequence and the second privacy sequence respectively held by the two business parties; wherein the first privacy sequence and the second privacy sequence are sorted according to a specified sorting direction; cooperating with the other business party of the two business parties to perform element filling processing on the initial auxiliary sequence to obtain an auxiliary sequence; wherein the element filling processing comprises: filling each element of the first privacy sequence into the initial auxiliary sequence from front to back, filling each element of the second privacy sequence into the initial auxiliary sequence from back to front, and configuring a reference element corresponding to the specified sorting direction at a position without an element; cooperating with the other business party of the two business parties to perform the following iterative steps on the auxiliary sequence: using the auxiliary sequence as the input sequence of the first round, performing splitting processing and position swapping processing; and then performing the splitting processing in the previous round. and the subsequence obtained after the position swapping process is used as the input sequence of this round, and the splitting process and the position swapping process are performed; it is iterated in sequence until the sequence length of the input sequence is 1, and the position information of each element in the auxiliary sequence in the last round is output; wherein the splitting process includes: splitting the input sequence from the middle to obtain two subsequences; the two subsequences include the front sequence segment and the back sequence segment of the input sequence, and the length of the front sequence segment is the same as that of the back sequence segment; the position swapping process includes: extracting elements from the front sequence segment and the back sequence segment in sequence, and comparing the elements extracted from the front sequence segment with the elements extracted from the back sequence segment; when the elements extracted from the front sequence segment and the elements extracted from the back sequence segment do not meet the specified sorting direction, swapping the positions of the elements extracted from the front sequence segment and the elements extracted from the back sequence segment; based on the position information of each element in the auxiliary sequence in the last round, extracting the sorted position information of each element in the first privacy sequence or the second privacy sequence held; and the other business party of the two business parties performs business collaborative processing according to the sorted position information of each element in the first privacy sequence and the second privacy sequence.
[0006] In some other embodiments, any round in the iterative steps is taken as the current round. Before performing splitting processing and position interchange processing on the input sequence of the current round, the method further includes: calculating the ratio of the length of the merged sequence to the sequence length of the input sequence of the current round, and rounding up the ratio to obtain the sorting sequence value of the current round; based on the order from front to back, extracting the input sequence corresponding to the number of the sorting sequence value from the input sequence of the auxiliary sequence in the current round as the sorting input sequence of the current round; in the current round, performing splitting processing and position interchange processing on the sorting input sequence.
[0007] In some other embodiments, before constructing the initial auxiliary sequence, the method further includes: sorting the first initial plaintext sequence or the second initial plaintext sequence held locally in a specified sorting direction and performing ciphertext processing to obtain the first privacy sequence or the second privacy sequence.
[0008] In some other embodiments, the sequence length of the initial auxiliary sequence is greater than or equal to the first power of 2 of the combined sequence length.
[0009] In some other embodiments, the specified sorting direction is a sorting direction from small to large; accordingly, the reference element is an infinite element.
[0010] In some other embodiments, the specified sorting direction is a sorting direction from large to small; accordingly, the reference element is an infinitesimal parameter.
[0011] On the other hand, the present specification also provides a business collaborative processing device based on privacy protection, which is applied to any business party of two business parties holding a sequence to be sorted in business collaborative processing, and the device includes: a sequence collaborative construction module, which is used to collaborate with the other business party of the two business parties to construct an initial auxiliary sequence based on the merged sequence length; wherein the sequence length of the initial auxiliary sequence is a value greater than or equal to the merged sequence length and is a power of 2; the merged sequence length is the sum of the lengths of the first privacy sequence and the second privacy sequence respectively held by the two business parties; wherein the first privacy sequence and the second privacy sequence are sorted according to a specified sorting direction; an element collaborative filling module, which is used to collaborate with the other business party of the two business parties to perform element filling processing on the initial auxiliary sequence to obtain an auxiliary sequence; wherein the element filling processing includes: filling each element of the first privacy sequence into the initial auxiliary sequence from front to back, filling each element of the second privacy sequence into the initial auxiliary sequence from back to front, and configuring a reference element corresponding to the specified sorting direction at a position without an element; an element position collaborative sorting module, which is used to collaborate with the other business party of the two business parties to perform the following iterative steps on the auxiliary sequence: using the auxiliary sequence as the input sequence of the first round, performing splitting processing and position sorting. The method comprises the following steps: taking the subsequence obtained after the previous round of splitting and position swapping as the input sequence of this round, performing splitting and position swapping; iterating in sequence until the sequence length of the input sequence is 1, and outputting the position information of each element in the auxiliary sequence in the last round; wherein the splitting process comprises: splitting the input sequence from the middle to obtain two subsequences; the two subsequences comprise a front sequence segment and a rear sequence segment of the input sequence, and the length of the front sequence segment is the same as that of the rear sequence segment; the position swapping process comprises: extracting elements from the front sequence segment and the rear sequence segment in sequence, and replacing the elements extracted from the front sequence segment with the elements extracted from the front sequence segment. The elements extracted from the front sequence segment are compared with the elements extracted from the rear sequence segment; when the elements extracted from the front sequence segment and the elements extracted from the rear sequence segment do not satisfy the specified sorting direction, the positions of the elements extracted from the front sequence segment and the elements extracted from the rear sequence segment are swapped; a position information extraction module is used to extract the sorted position information of each element in the first privacy sequence or the second privacy sequence held based on the position information of each element in the auxiliary sequence in the last round; a business collaborative processing module is used to collaborate with the other business party of the two business parties to perform business collaborative processing according to the sorted position information of each element in the first privacy sequence and the second privacy sequence.
[0012] In some other embodiments, the device also includes a sorting module and a ciphertext processing module; wherein the sorting module is used to locally sort the first initial plaintext sequence or the second initial plaintext sequence held in a specified sorting direction to obtain a first plaintext sequence or a second plaintext sequence; the ciphertext processing module is used to ciphertext process the first plaintext sequence or the second plaintext sequence held to obtain a first privacy sequence or a second privacy sequence.
[0013] On the other hand, an embodiment of the present specification further provides a computer-readable storage medium having computer instructions stored thereon, and when the instructions are executed by a processor, the steps of the method described in any one or more of the above embodiments are implemented.
[0014] On the other hand, an embodiment of the present specification also provides a computer device, including a memory, a processor, and computer instructions stored in the memory, wherein the processor executes the computer instructions to implement the steps of the method described in any one or more of the above embodiments.
[0015] The business collaborative processing method, device, medium and equipment based on privacy protection provided by one or more embodiments of this specification can protect the data privacy of the business party as much as possible, so that multiple business parties can realize business collaborative processing based on privacy protection. At the same time, the initial plaintext sorting is performed locally, and no communication between nodes is required. Compared with ciphertext sorting, the time consumption of plaintext sorting is negligible; and the collaborative sorting under ciphertext greatly reduces the sorting rounds, and is more suitable for parallel processing, which can greatly reduce the complexity of communication and data processing, making the overall business collaborative processing simpler and more efficient. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] In order to more clearly illustrate the technical solutions in the embodiments of this specification or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art description. Obviously, the drawings described below are only some embodiments recorded in this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. In the drawings:
[0017] Figure 1 A schematic diagram of the business collaborative processing interaction process based on privacy protection provided in this manual;
[0018] Figure 2 A local sequence order and auxiliary sequence diagram provided for this specification;
[0019] Figure 3 A schematic diagram of the sequence co-sequencing provided in this specification;
[0020] Figure 4A schematic diagram of the implementation process of the business collaborative processing method based on privacy protection provided in this specification;
[0021] Figure 5 A schematic diagram of the business collaborative processing module structure based on privacy protection provided in this manual;
[0022] Figure 6 This is a schematic diagram of the module structure of the computer device provided in this manual. DETAILED DESCRIPTION
[0023] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in one or more embodiments of this specification will be clearly and completely described below in conjunction with the drawings in one or more embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of the specification, not all of the embodiments. Based on one or more embodiments of the specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the embodiments of this specification.
[0024] The solutions provided in the embodiments of this specification can be applied to various business parties in business collaborative processing. The various business parties in the business collaborative processing may be devices corresponding to enterprises, units, individuals, etc. In view of the differences in the business parties involved in the business collaborative processing, the plaintext data of each business party involved in the business collaborative processing process needs to be privacy protected and cannot be directly exposed to other business parties during the business collaborative processing process, so as to achieve the purpose of privacy protection of the business data held by each business party. Under this business application requirement, the business collaborative processing based on privacy protection may include: model collaborative training and collaborative prediction based on privacy protection, business collaborative data analysis based on privacy protection, etc.
[0025] In a scenario example, Figure 1 As shown, assuming that the business parties of business collaborative processing are P1 and P2, each business party can read business data from the corresponding business system respectively, and perform business collaborative processing based on multi-party secure computing. Of course, the process of business collaborative processing can also involve a coordination service device, which can communicate with the devices of each business party to execute or unify the processing instructions involved in the process of business collaborative processing.
[0026] In the business processing processes such as collaborative model training and collaborative prediction based on privacy protection, and business collaborative data analysis based on privacy protection, the collaborative sorting of business data of each business party is one of the business processing sub-links that is usually involved in many business collaborative processing processes. For example, data binning processing involves multi-party collaborative sorting of feature data. Therefore, how to process the collaborative sorting of business data based on privacy protection is very critical to business collaborative processing. In the process of collaborative sorting of business data based on privacy protection, it is necessary to collaboratively sort the business data without disclosing the business data held by each business party, so as to protect the privacy of the business data involved in business collaborative processing as much as possible.
[0027] For example, in the process of executing business collaborative processing, it usually involves business party P1 holding business data of some business types and business party P2 holding business data of another business type. It is necessary to comprehensively utilize the business data held by the two business parties to collaboratively complete the business processing.
[0028] For example, in the process of user risk prediction, it usually involves the situation that business party P1 holds user commodity transaction data and business party P2 holds basic information of user accounts. In the process of risk prediction, it is necessary to integrate user commodity transaction data and basic information of user accounts to achieve user risk prediction. However, in the process of user risk prediction based on privacy protection, it is necessary to ensure that during the user risk prediction process, the user commodity transaction data held by business party P1 and the basic information of user accounts held by business party P2 do not leave the local devices of each business party, and there is no risk of indirect leakage, so as to protect the privacy of user data to the greatest extent possible. Accordingly, user risk prediction needs to be completed by the two business parties in collaboration under privacy protection.
[0029] In the risk prediction process, the construction of the risk prediction model and the execution of risk prediction based on the risk prediction model usually involve the process of sorting the feature vectors of the input model, such as sorting the feature vectors from small to large, or from large to small. Given that the feature vectors are jointly held by two business parties and need to be privacy protected, it is very important for collaborative risk prediction to quickly and easily perform collaborative sorting on the feature vector shards held by the two parties without leaking the feature vector shard information held by each party, so as to achieve the overall sorting of the feature vector.
[0030] like Figure 2As shown, in this scenario example, sequence X1 is an example of a feature vector slice held by business party P1, and sequence Y1 is an example of a feature vector slice held by business party P2. Business parties P1 and P2 need to collaborate to sort arrays X1 and Y1, such as sorting the elements involved in X1 and Y1 from small to large, or from large to small. In this scenario example, the following is an example of sorting the elements involved in X1 and Y1 from small to large.
[0031] Business parties P1 and P2 can first sort sequences X1 and Y1 locally to obtain X2 and Y2. Since sequences X1 and Y1 are held by business parties P1 and P2 respectively, they can be quickly executed in plain text when sorting locally.
[0032] After the local sorting is completed, the business parties P1 and P2 can also perform ciphertext processing on the sequences X2 and Y2 locally respectively to obtain the ciphertext sequences X3 and Y3. Afterwards, the business parties P1 and P2 can perform collaborative sorting processing based on the ciphertext sequence to avoid leakage of business information held by each business party during the collaborative processing. For example, the sequences X2 and Y2 can be ciphertext processed using a homomorphic encryption algorithm so that the result of a certain operation on the ciphertext data is exactly equal to the ciphertext obtained after the expected operation is performed on the plaintext data before encryption. In this scenario example, for the sake of intuitive explanation, the ciphertext sequence is still displayed in plaintext because its plaintext form is the sequence X2 and Y2. Therefore, Figure 2 The elements in the ciphertext sequences X3 and Y3 are the same as those in the sequences X2 and Y2. It should be noted that in the actual application scenario, the ciphertext forms of the sequences X2 and Y2 are collaboratively sorted.
[0033] In the collaborative sorting process, business parties P1 and P2 can first collaboratively calculate the sum of the lengths of sequences X3 and Y3. For ease of description, the sum of the lengths of sequences X3 and Y3 can be used as the merged sequence length. Then, business parties P1 and P2 can collaboratively construct an initial auxiliary sequence based on the merged sequence length. Among them, the sequence length of the initial auxiliary sequence can be set to a value that is greater than or equal to the merged sequence length and is a power of 2. Preferably, the first power of 2 that is greater than or equal to the merged sequence length can be set as the sequence length of the initial auxiliary sequence to further reduce the amount of data processing and interaction complexity involved in the collaborative sorting process, and improve the processing efficiency of the collaborative sorting.
[0034] Afterwards, the service parties P1 and P2 may collaboratively perform the following element filling steps on the initial auxiliary sequence to obtain an auxiliary sequence.
[0035] For example, business party P1 can fill each element of X3 into the initial auxiliary sequence locally from front to back, and business party P2 can fill each element of Y3 into the initial auxiliary sequence locally from back to front. Business parties P1 and P2 can determine whether there are any positions with extra unfilled elements after the initial auxiliary sequence has been filled with elements. For example, business parties P1 and P2 can determine whether the sequence length of the initial auxiliary sequence is equal to the merged sequence length. If not, it can be said that after the initial auxiliary sequence has been filled with elements, there are positions with extra unfilled elements. In the case where there are positions with extra unfilled elements, business parties P1 and P2 can determine the positions that need to be filled with infinite elements based on the sequence lengths of X3 and Y3 and the filling direction, and fill in infinite elements at the corresponding positions to obtain an auxiliary sequence.
[0036] Alternatively, business parties P1 and P2 may first determine whether the sequence length of the initial auxiliary sequence is equal to the length of the merged sequence. If not, it may indicate that after the initial auxiliary sequence is filled with elements, there are positions with extra unfilled elements. In the case where there are positions with extra unfilled elements, business parties P1 and P2 may determine the positions that need to be filled with infinite elements based on the sequence lengths of X3 and Y3 and the filling direction, and fill in infinite elements at the corresponding positions. Afterwards, business party P1 may fill each element of X3 into the initial auxiliary sequence locally from front to back, and business party P2 may fill each element of Y3 into the initial auxiliary sequence locally from back to front to obtain an auxiliary sequence.
[0037] like Figure 2 As shown, Z is an auxiliary sequence, wherein the first 9 bits of Z are filled with the elements of sequence X3 from the front to the back, and the last 2 bits of Z are filled with the elements of sequence Y3 from the back to the front. There is an infinite element (M) after the last element of sequence X3 and before the first element of sequence Y3.
[0038] Of course, business parties P1 and P2 can also use other variations to fill in elements and obtain auxiliary sequences. Since X3 and Y3 are arranged from small to large, the elements of sequence Y3 can be filled in from front to back, and the elements of sequence X3 can be filled in from back to front.
[0039] Since sequence X3 is held locally by business party P1 and Y3 is held locally by business party P2, the elements held by business party P1 in auxiliary sequence Z are at most elements of sequence X3 and infinite elements, and the elements held by business party P2 in auxiliary sequence Z are at most elements of sequence Y3 and infinite elements. Accordingly, business parties P1 and P2 need to collaborate to complete the overall sorting of auxiliary sequence Z.
[0040] In this scenario example, business parties P1 and P2 can collaboratively perform the following iterative steps on the auxiliary sequence to collaboratively complete the overall sorting of the auxiliary sequence:
[0041] The auxiliary sequence is used as the input sequence of the first round, and the splitting process and the position exchange process are performed.
[0042] The splitting process may include: splitting the input sequence from the middle to obtain two subsequences; the two subsequences include a front sequence segment and a rear sequence segment of the input sequence, and the front sequence segment has the same length as the rear sequence segment.
[0043] The position swapping process may include: extracting elements from the preceding sequence segment and the succeeding sequence segment in sequence, and comparing the elements extracted from the preceding sequence segment with the elements extracted from the succeeding sequence segment; and swapping the positions of the elements extracted from the preceding sequence segment with the elements extracted from the succeeding sequence segment when the elements extracted from the preceding sequence segment and the elements extracted from the succeeding sequence segment do not satisfy the specified sorting direction.
[0044] The subsequence obtained after the previous round of splitting and position exchange processing is used as the input sequence of this round, and the splitting processing and position exchange processing are performed.
[0045] Iterate in sequence until the sequence length of the input sequence is 1, and output the position information of each element in the auxiliary sequence in the last round.
[0046] For example, Figure 3 As shown, the auxiliary sequence Z can be used as the input sequence of the first round and split from the middle to obtain the front sequence segment Z 11 And the subsequent sequence Z 12 Two subsequences. From the previous sequence segment Z 11 And the subsequent sequence Z 12 Extract the elements in sequence and put them from the previous sequence segment Z 11 The extracted elements are the same as those from the subsequent sequence segment Z 12 The extracted elements are compared. 11 Extract the first element 0 from the previous sequence segment Z 12 The first element 15 extracted is compared. If 0 is smaller than 15, the positions of elements 0 and 15 are kept unchanged. 11 The second element 2 extracted from the previous sequence segment Z 12 The second element M extracted is compared. If 2 is smaller than M, the position of element 2 and element M remains unchanged. Similarly, the extracted elements are compared. 11 The extracted element is smaller than the one from the previous sequence segment Z 12 If the extracted elements are small or equal, keep the position of the extracted elements unchanged; in the previous sequence segment Z11 The extracted element is smaller than the one from the previous sequence segment Z 12 When the extracted element is large, swap the position of the extracted element (such as Z 11 The element 14 and Z 12 Elements 3).
[0047] When business parties P1 and P2 collaborate to perform the above element extraction, the element extraction can be implemented using a one-hot vector to avoid exposing the position information of the element in the sorting. Of course, the element extraction can also be implemented using other privacy protection processing methods, which are not limited in this specification.
[0048] In actual operation, the extracted Z 11 With Z 12 The first element of is actually held by business party P1. At this time, business party P1 can compare the two extracted elements locally. Since there is no need to swap positions, business party P1 records the initial position information of the two elements "0" and "15" in this round as the position information after the position swap in this round. 11 With Z 12 The last two elements of are held by business parties P1 and P2 respectively. At this time, the comparison of element sizes can be realized by business parties P1 and P2 using private_compare under MPC. If there is a position swap, business parties P1 and P2 respectively record the position information of the elements they hold after the swap as the position information after the position swap in this round; if there is no position swap, business parties P1 and P2 respectively record the initial position information of the elements they hold in this round as the position information after the position swap in this round. The extracted Z 11 With Z 12 As for other elements of , since they involve the infinite element M, it is impossible to involve position swapping. Business party P1 records the initial position information of the elements it holds and the infinite element M in this round, and business party P2 records the initial position information of the infinite element M in this round as the position information after the position swap in this round.
[0049] Through the above method, business parties P1 and P2 can work together to complete Z 11 With Z 12 The positions of the elements in the array are swapped, and the position information of the elements (including infinite elements) held by the array is recorded after the position swap. For the convenience of expression and display, Figure 3 As shown, after the element positions are swapped, the previous sequence segment Z′ is obtained in the form 11 And the subsequent sequence segment Z′ 12 .
[0050] By comparing Z′ 11 , Z′12 It can be seen that Z′ 11 All elements in are smaller than Z′ 12 That is, in the above round of position swapping, all the smaller elements are adjusted to the front sequence segment, and all the larger elements are adjusted to the back sequence segment.
[0051] Afterwards, a second round of iterative steps can be performed, with Z′ 11 , Z′ 12 As input sequences respectively. For example, for Z′ 11 Perform the splitting process to get the previous sequence segment Z 111 And the subsequent sequence Z 112 Two subsequences. You can start from the previous sequence segment Z 111 And the subsequent sequence Z 112 Extract the elements one by one and put them from the previous sequence segment Z 111 The extracted elements are the same as those from the subsequent sequence segment Z 112 The extracted elements are compared. In the previous sequence segment Z 111 The extracted element is smaller than the one from the previous sequence segment Z 112 If the extracted elements are small or equal, keep the position of the extracted elements unchanged; in the previous sequence segment Z 111 The extracted element is smaller than the one from the previous sequence segment Z 112 When the extracted element is large, swap the positions of the extracted elements to get Z′ 111 , Z′ 112 . And, Z′ 12 Perform splitting processing to obtain the front sequence segment Z 121 And the subsequent sequence Z 122 Two subsequences. And the previous sequence segment Z 111 And the subsequent sequence Z 112 Perform the same position swapping process on each element of Z to get Z′ 121 , Z′ 122 , I will not go into details here.
[0052] Afterwards, a third round of iterative steps can be performed, with Z′ 111 , Z′ 112 , Z′ 121 , Z′ 122 As input sequences, splitting and position swapping are performed respectively. And so on, until the length of the subsequence finally split is 1. For example, in the fourth round of Z in this scenario example, the length of the subsequence split is 1. If the fifth round is executed, the length of the input sequence of the fifth round is 1, and the splitting process cannot be performed anymore. Therefore, there is no need to perform the fifth round, and the fourth round is the last round of the iterative step. After the fourth round is executed, the position information of each element after the fourth round is completed can be output.
[0053] In the above iterative steps, each round adjusts the smaller elements to the front sequence segment of the input sequence and the larger elements to the back sequence segment, thereby realizing the overall sorting of the elements from small to large, and obtaining the auxiliary sequence Z' in which the elements are arranged from small to large. Correspondingly, the first 11 elements in the auxiliary sequence Z' are the elements in the sequences X3 and Y3 after being sorted from small to large.
[0054] Of course, the above auxiliary sequence Z, the subsequences of each round, and the auxiliary sequence Z ′ in which the elements are arranged from small to large are only examples for the convenience of intuitive explanation. In actual application scenarios, the elements of the sequences X3 and Y3 held by the business parties P1 and P2 will not leave the business parties P1 and P2. The business party P1 finally obtains the position information of each element in X3 and the infinite element in the last round, and the business party P2 finally obtains the position information of each element in Y3 and the infinite element in the last round. The infinite element is only an auxiliary element, and all are arranged at the end, which has no effect on the sorting of all elements involved in X3 and Y3 from small to large. Therefore, the business party P1 can extract the position information of each element in X3 in the last round as the position information of each element in X3 after sorting; the business party P2 can extract the position information of each element in Y3 in the last round as the position information of each element in Y3 after sorting, so that the business parties P1 and P2 are equivalent to realizing the sorting of each element involved in X3 and Y3 as a whole.
[0055] Taking any round in the iteration step as the current round, before performing the splitting process and the position swapping process on the input sequence of the current round, a sequence extraction process may also be performed. The sequence extraction process may include calculating the ratio of the length of the merged sequence to the sequence length of the input sequence of the current round, and rounding up the ratio to obtain the sorting sequence value of the current round; extracting the input sequence corresponding to the sorting sequence value from the input sequence of the auxiliary sequence in the current round based on the order from front to back, as the sorting input sequence of the current round; performing the splitting process and the position swapping process on the sorting input sequence in the current round.
[0056] Since a large number of infinite elements may be added to the auxiliary sequence, even if the first power of 2 of the merged sequence length is used as the sequence length of the auxiliary sequence, the number of infinite elements may be close to the total number of elements held by the two business parties (for example, when the merged sequence length is 9, the sequence length of the auxiliary sequence is 16, and correspondingly, the number of infinite elements is 7, which nearly doubles the amount of data required for collaborative sorting), which leads to a significant increase in the amount of interactive transmission data and the complexity of interaction of the business party. In the above-mentioned iterative process, infinite elements are gradually placed in the back position, so that a certain input sequence may be entirely infinite elements. At this time, there is no need to perform splitting and position swapping on the input sequence. Through the above-mentioned sequence extraction process, the input sequence whose elements are all infinite elements can be eliminated, and the splitting and position swapping processes are no longer performed, which can further effectively reduce the amount of interactive transmission data and the complexity of interaction of the business party during the iteration process. For example, when the combined sequence length is 9, if the sequence extraction process is not performed, the input sequence that needs to be split and interchanged in the third round is 4, and the fourth round is 8; if the sequence extraction process is performed, the input sequence that needs to be split and interchanged in the second round is 3, and the third round is only 5.
[0057] In another scenario example, the elements involved in the sequences X1 and Y1 can also be sorted from large to small. As in the above scenario example, the business parties P1 and P2 can first sort the sequences X1 and Y1 locally, and perform ciphertext processing to obtain two ciphertext sequences in which the elements are sorted from large to small. Then, the business parties P1 and P2 perform collaborative sorting processing on the two ciphertext sequences. Among them, when constructing the auxiliary sequence, if there are positions with extra unfilled elements, the business parties P1 and P2 can fill in the positions where the infinite elements are filled in based on the solution of the above scenario example; the rest is the same as the above scenario example.
[0058] During the iteration process, the position swapping process may include: extracting elements from the front sequence segment and the rear sequence segment in sequence, and comparing the elements extracted from the front sequence segment with the elements extracted from the rear sequence segment; when the elements extracted from the front sequence segment are smaller than the elements extracted from the rear sequence segment, swapping the positions of the elements extracted from the front sequence segment with the elements extracted from the rear sequence segment. Other processing during the iteration process may be the same as the above scenario example. Through this iteration process, each round can adjust the larger elements to the front sequence segment of the input sequence and the smaller elements to the rear sequence segment, thereby realizing the overall sorting of the elements from large to small, and all infinitesimal elements are placed at the end. Afterwards, the business parties P1 and P2 can determine the sorted position information of each element in the ciphertext sequence they hold based on the position information of each element in the auxiliary sequence in the last round.
[0059] If business collaborative processing involves multiple business parties, for example, business parties P0, P1, P2, P3, and P4 have sequences U, V, W, X, and Y respectively, P0 and P1 can collaboratively sort sequence U and sequence V to obtain sequence Q; P2 and P3 can collaboratively sort sequence W and sequence X to obtain sequence R; then, P0 and P1 can be regarded as one business party, and P2 and P3 can be regarded as one business party, and P0, P1, P2, and P3 can collaboratively sort sequence Q and sequence R to obtain sequence S; finally, P0, P1, P2, and P3 can be regarded as one business party, and collaboratively sort sequence S and sequence Y with business party P4 to obtain sequence T, thereby completing the collaborative sorting of multiple business parties based on the implementation method given in the above scenario example. Of course, the above sequences Q, R, S, and T are sequences constructed based on the position information of each element after sorting for the sake of convenience. In actual application scenarios, each required element is still held by each business party, and only the position information of the element is adjusted so that the elements held by each business party are formally sorted as a whole.
[0060] It should be noted that the above-mentioned combination is only a preferred example. The specific embodiment may also adopt other combinations, such as P1 and P2 as a group, P3 and P4 as a group; or, P0 and P1 as a group, and then P0, P1, and P3 as a group, etc. However, as long as the functions and effects achieved are the same or similar, they should be covered within the scope of protection of this specification.
[0061] The implementation method provided by the above scenario example gives a secure collaborative sorting scheme under MPC, which can protect the data privacy of the business party as much as possible and enable multiple business parties to implement business collaborative processing based on privacy protection. At the same time, the initial plaintext sorting is performed locally, and no communication between nodes is required. Compared with ciphertext sorting, the time consumption of plaintext sorting is negligible. The communication round complexity involved in the collaborative sorting under ciphertext is O(log2(n)), and the computational complexity is only O(n*log2(n)), where n is the number of elements in the auxiliary sequence. It can be seen that although the number of elements involved in the collaborative sorting process under ciphertext may be more than the number of elements actually to be sorted by the business party, the above collaborative sorting method is relative to the method of comparing elements one by one between the two business parties. The sorting rounds are still greatly reduced, and it is more suitable for parallel processing, which can greatly reduce the complexity of communication and data processing, making the overall business collaborative processing simpler and more efficient.
[0062] Based on the above scenario examples, the embodiments of this specification provide a business collaborative processing method based on privacy protection, which can be applied to any business party of the two business parties holding the sequence to be sorted in business collaborative processing. For example, when the sequences to be sorted are U and V, the two business parties are P0 and P1. When the sequences to be sorted are sequence Q and sequence R, the two business parties are (P0, P1), (P2, P3), and so on. Figure 4 As shown, the method may include the following steps.
[0063] S40: Collaborating with the other service party of the two service parties to construct an initial auxiliary sequence based on the combined sequence length; wherein the sequence length of the initial auxiliary sequence is a value greater than or equal to the combined sequence length and is a power of 2.
[0064] The business collaborative processing may be a business processing method collaboratively performed by at least two business parties. As described in the above scenario example, the business collaborative processing may be a user risk prediction collaboratively performed by business parties P1 and P2. Of course, in other scenario examples, business collaborative processing may also include other business processing types, such as data binning processing, etc. In the process of business collaborative processing based on privacy protection, the collaborative sorting of business data of each business party is one of the business processing sub-links that is usually involved in many business collaborative processing processes. Therefore, how to collaboratively sort business data based on privacy protection is very critical to business collaborative processing.
[0065] In various business processing processes, the business data to be processed can be stored in the form of a sequence. For example, in the above-mentioned user risk prediction scenario, feature data is usually stored in the form of a vector, namely a feature vector. Among them, each element of the feature vector corresponds to a different feature value. Accordingly, the feature vector can be a data sequence at the data level.
[0066] In order to facilitate the recording of the position information of the data sequence during the sorting process, a position identifier can be configured for each element position in the data sequence. For example, the position identifiers can be configured for each element position of the data sequence from left to right: 1, 2, 3, ... Of course, the position identifiers can also be configured for each element position of the data sequence from left to right: 0, 1, 2, ...; or, the position identifiers can also be configured for each element position of the data sequence from right to left: 0, 1, 2, ...; of course, the position identifiers of each element position in the data sequence may not increase from left to right or from right to left, but may be randomly determined or given based on other rules, which is not limited here. Of course, in order to facilitate the identification of elements during the sorting process, element identifiers can also be configured for each element. Alternatively, other methods can be used to identify the position information of each element position of the data sequence and to identify each element, which is not limited in this specification.
[0067] In the embodiments of this specification, for ease of description, the sequences to be sorted held by the two business parties can be described as the first initial plaintext sequence and the second initial plaintext sequence, respectively. As described in the above scenario example, the sequences to be sorted are feature vector fragments, that is, the first initial plaintext sequence and the second initial plaintext sequence are feature vector fragments held by each business party, respectively.
[0068] The two business parties can respectively sort the first initial plaintext sequence and the second initial plaintext sequence held locally in a specified sorting direction to obtain the first plaintext sequence and the second plaintext sequence. Correspondingly, each element in the first plaintext sequence and the second plaintext sequence is sorted in the specified sorting direction. The specified sorting direction can be sorting from small to large or from large to small.
[0069] The two business parties can respectively perform ciphertext processing on the first plaintext sequence and the second plaintext sequence they hold locally to obtain the first privacy sequence and the second privacy sequence, so as to further avoid leakage of the business information held by the business parties during the collaborative sorting process.
[0070] The two business parties can collaboratively calculate the sum of the lengths of the first privacy sequence and the second privacy sequence. For ease of description, the sum of the lengths of the first privacy sequence and the second privacy sequence can be described as the combined sequence length. Then, the two business parties can collaboratively construct an initial auxiliary sequence based on the combined sequence length. For example, a value that is greater than or equal to the combined sequence length and is a power of 2 can be extracted as the sequence length of the initial auxiliary sequence. Preferably, the first power of 2 that is greater than or equal to the combined sequence length can be extracted as the sequence length of the initial auxiliary sequence.
[0071] S42: Collaborating with another service party of the two service parties to perform element filling processing on the initial auxiliary sequence to obtain an auxiliary sequence.
[0072] The two business parties may collaborate to perform element filling processing on the initial auxiliary sequence to obtain an auxiliary sequence. The element filling processing may include: filling each element of the first privacy sequence into the initial auxiliary sequence from front to back, and filling each element of the second privacy sequence into the initial auxiliary sequence from back to front, and configuring the reference element corresponding to the specified sorting direction at the position where there is no element. In the case where the specified sorting direction is a sorting direction from small to large, the reference element is an infinite element. In the case where the specified sorting direction is a sorting direction from large to small, the reference element is an infinitesimal element. The implementation method of the element filling processing can be referred to the above-mentioned scenario example, which will not be repeated here.
[0073] S44: Cooperate with the other business party of the two business parties to perform iterative steps on the auxiliary sequence: use the auxiliary sequence as the input sequence of the first round, perform splitting processing and position exchange processing; then use the subsequence obtained after the previous round of splitting and position exchange processing as the input sequence of this round, perform splitting processing and position exchange processing; iterate in sequence until the sequence length of the input sequence is 1, and output the position information of each element in the auxiliary sequence in the last round.
[0074] Two business parties can collaborate to perform iterative steps on the auxiliary sequence:
[0075] The auxiliary sequence is used as the input sequence of the first round to perform splitting and position exchange processing.
[0076] The splitting process may include: splitting the input sequence from the middle to obtain two subsequences; the two subsequences include a front sequence segment and a rear sequence segment of the input sequence, and the front sequence segment has the same length as the rear sequence segment.
[0077] The position swapping process may include: extracting elements from the preceding sequence segment and the succeeding sequence segment in sequence, and comparing the elements extracted from the preceding sequence segment with the elements extracted from the succeeding sequence segment; and swapping the positions of the elements extracted from the preceding sequence segment with the elements extracted from the succeeding sequence segment when the elements extracted from the preceding sequence segment and the elements extracted from the succeeding sequence segment do not satisfy the specified sorting direction.
[0078] The fact that the elements extracted from the front sequence segment and the elements extracted from the rear sequence segment do not satisfy the specified sorting direction may mean that: when the specified sorting direction is from small to large, according to the specified sorting direction, the front element should be less than or equal to the rear element, therefore, the element extracted from the front sequence segment is greater than the element extracted from the rear sequence segment, i.e., the specified sorting direction is not satisfied; and when the specified sorting direction is from large to small, according to the specified sorting direction, the front element should be greater than or equal to the rear element, therefore, the element extracted from the front sequence segment is less than the element extracted from the rear sequence segment, i.e., the specified sorting direction is not satisfied.
[0079] The subsequence obtained after the previous round of splitting and position exchange processing is used as the input sequence of this round, and the splitting processing and position exchange processing are performed.
[0080] Iterate in sequence until the sequence length of the input sequence is 1, and output the position information of each element in the auxiliary sequence in the last round.
[0081] The implementation method of the iterative process can refer to the above scenario example, which will not be described in detail here.
[0082] In the above-mentioned iterative process, in some other embodiments, any round in the iterative step is taken as the current round, and before the splitting process and the position exchange process are performed on the input sequence of the current round, the method may further include a sequence extraction process. The sequence extraction process may include calculating the ratio of the length of the merged sequence to the sequence length of the input sequence of the current round, and rounding up the ratio to obtain the sorting sequence value of the current round; based on the order from front to back, extracting the input sequence corresponding to the number of the sorting sequence value from the input sequence of the auxiliary sequence in the current round as the sorting input sequence of the current round; in the current round, performing the splitting process and the position exchange process on the sorting input sequence.
[0083] S46: Based on the position information of each element in the auxiliary sequence in the last round, extract the sorted position information of each element in the first privacy sequence or the second privacy sequence.
[0084] S48: Another business party that cooperates with the two business parties performs business collaborative processing according to the position information of the elements in the first privacy sequence and the second privacy sequence after being sorted.
[0085] After each business party obtains the position information of each element in the privacy sequence after sorting, each business party can use the position information of each element in the privacy sequence after sorting to perform other business processing links of business collaborative processing under privacy protection, such as the model construction of user risk prediction and other links of actual prediction business processing.
[0086] The implementation methods provided by the above embodiments can protect the data privacy of the business parties as much as possible, and enable multiple business parties to achieve business collaborative processing based on privacy protection. At the same time, the initial plaintext sorting is performed locally, and no communication between nodes is required. Compared with ciphertext sorting, the time consumption of plaintext sorting can be ignored; and the collaborative sorting under ciphertext greatly reduces the sorting rounds, and is more suitable for parallel processing, which can greatly reduce the complexity of communication and data processing, making the overall business collaborative processing simpler and more efficient.
[0087] Of course, inspired by the technical essence of the embodiments of this specification, technicians in the relevant field may also come up with other modified implementation methods, but as long as the functions and effects achieved are the same or similar, they should be included in the protection scope of this specification.
[0088] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. For details, please refer to the description of the above-mentioned related processing related embodiments, and no further description is given here.
[0089] The above is a description of a specific embodiment of the specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0090] like Figure 5 As shown, based on the above-mentioned business collaborative processing method based on privacy protection, the embodiment of this specification also provides a business collaborative processing device based on privacy protection, which is applied to any business party of two business parties holding sequences to be sorted in business collaborative processing, and the device includes:
[0091] The sequence collaborative construction module 502 is used to collaborate with the other business party of the two business parties to construct an initial auxiliary sequence based on the combined sequence length; wherein the sequence length of the initial auxiliary sequence is a value greater than or equal to the combined sequence length and a power of 2; the combined sequence length is the sum of the lengths of the first privacy sequence and the second privacy sequence respectively held by the two business parties; wherein the first privacy sequence and the second privacy sequence are sorted in a specified sorting direction;
[0092] The element collaborative filling module 504 is used to collaborate with the other business party of the two business parties to perform element filling processing on the initial auxiliary sequence to obtain an auxiliary sequence; wherein the element filling processing includes: filling each element of the first privacy sequence into the initial auxiliary sequence from front to back, filling each element of the second privacy sequence into the initial auxiliary sequence from back to front, and configuring the reference element corresponding to the specified sorting direction at the position without the element;
[0093] The element position collaborative sorting module 506 is used to collaborate with the other business party of the two business parties to perform the following iterative steps on the auxiliary sequence: using the auxiliary sequence as the input sequence of the first round, performing splitting processing and position swapping processing; then using the subsequence obtained after the previous round of splitting and position swapping processing as the input sequence of this round, performing splitting processing and position swapping processing; iterating in sequence until the sequence length of the input sequence is 1, and outputting the position information of each element in the auxiliary sequence in the last round; wherein the splitting processing includes: splitting the input sequence from the middle to obtain two subsequences; the two subsequences include the front sequence segment and the rear sequence segment of the input sequence, and the length of the front sequence segment is the same as that of the rear sequence segment; the position swapping processing includes: extracting elements from the front sequence segment and the rear sequence segment in sequence, and comparing the elements extracted from the front sequence segment with the elements extracted from the rear sequence segment; when the elements extracted from the front sequence segment and the elements extracted from the rear sequence segment do not satisfy the specified sorting direction, swapping the positions of the elements extracted from the front sequence segment with the elements extracted from the rear sequence segment;
[0094] A position information extraction module 508, configured to extract the sorted position information of each element in the first privacy sequence or the second privacy sequence based on the position information of each element in the auxiliary sequence in the last round;
[0095] The business collaborative processing module 510 is used to collaborate with another business party of the two business parties to perform business collaborative processing according to the position information of the elements in the first privacy sequence and the second privacy sequence after sorting.
[0096] In some other embodiments, the device also includes a sorting module and a ciphertext processing module; wherein the sorting module is used to locally sort the first initial plaintext sequence or the second initial plaintext sequence held in a specified sorting direction to obtain a first plaintext sequence or a second plaintext sequence; the ciphertext processing module is used to ciphertext process the first plaintext sequence or the second plaintext sequence held to obtain a first privacy sequence or a second privacy sequence.
[0097] It should be noted that the above-mentioned device may also include other implementations according to the description of the method embodiment and the scenario example. The specific implementation methods can refer to the description of the relevant method embodiment, and will not be described one by one here.
[0098] It should be noted that the above-mentioned device may also include other implementations according to the description of the method embodiment and the scenario example. The specific implementation methods can refer to the description of the relevant method embodiment, and will not be described one by one here.
[0099] like Figure 6As shown, based on the above-mentioned privacy protection-based collaborative business processing method, this specification also provides a computer device, the computer device includes at least one processor and a memory for storing processor executable instructions, and the instructions are executed by the processor to implement the steps of the method described in any one or more of the above embodiments. The processor can take the form of a microprocessor or processor and a computer-readable medium that stores a computer-readable program code (such as software or firmware) that can be executed by the (micro) processor, a logic gate, a switch, an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), a programmable logic controller, and an embedded microcontroller, etc. The memory includes but is not limited to a random access memory (Random Access Memory, RAM), a read-only memory (Read-Only Memory, ROM), a cache (Cache), a hard disk (Hard Disk Drive, HDD) or a memory card (MemoryCard). The computer device can be a server of the method described in the above embodiment, etc. Of course, the computer device can also include module structures such as input devices and transmission modules. In this embodiment, the functions and effects specifically implemented by the computer device can be explained in comparison with other embodiments, and will not be repeated here.
[0100] In addition, the embodiments of this specification also provide a computer-readable storage medium on which computer instructions are stored, and when the instructions are executed, the steps of the method described in any one or more of the above embodiments are implemented. The storage medium may include a physical device for storing information, which is usually to digitize the information and then store it in a medium using electrical, magnetic or optical methods. The storage medium may include: devices that store information using electrical energy, such as various types of memories, such as RAM, ROM, etc.; devices that store information using magnetic energy, such as hard disks, floppy disks, magnetic tapes, magnetic core memories, bubble memories, and U disks; devices that store information optically, such as CDs or DVDs. Of course, there are other types of readable storage media, such as quantum memories, graphene memories, and so on.
[0101] It should be noted that the embodiments of this specification are not limited to the situations that must comply with the standard data model / template or the embodiments of this specification. Certain industry standards or slightly modified implementation plans based on the implementation described in the custom method or the embodiment can also achieve the same, equivalent or similar, or predictable implementation effects after deformation of the above-mentioned embodiments. The embodiments obtained by using these modified or deformed data acquisition, storage, judgment, processing methods, etc. can still fall within the scope of the optional implementation plans of this specification.
[0102] Each embodiment in this specification is described in a progressive manner, and the same and similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. In the description of this specification, the description of the reference terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of this specification. In this specification, the schematic representation of the above terms does not necessarily target the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, in the absence of contradiction, a person skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.
[0103] The above description is only an embodiment of the present specification and is not intended to limit the present specification. For those skilled in the art, the present specification may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present specification shall be included in the scope of the claims of the present specification.
Claims
1. A business collaborative processing method based on privacy protection, characterized in that: Applied to any one of two business parties holding sequences to be sorted in business collaborative processing; the method comprises: Another business party cooperating with the two business parties constructs an initial auxiliary sequence based on the combined sequence length; wherein the sequence length of the initial auxiliary sequence is a value greater than or equal to the combined sequence length and is a power of 2; the combined sequence length is the sum of the lengths of the first privacy sequence and the second privacy sequence respectively held by the two business parties; wherein the first privacy sequence and the second privacy sequence are sorted in a specified sorting direction; The other business party of the two business parties performs element filling processing on the initial auxiliary sequence to obtain an auxiliary sequence; wherein the element filling processing includes: filling each element of the first privacy sequence into the initial auxiliary sequence from front to back, filling each element of the second privacy sequence into the initial auxiliary sequence from back to front, and configuring the reference element corresponding to the specified sorting direction at the position without the element; The other business party of the two business parties performs the following iterative steps on the auxiliary sequence: using the auxiliary sequence as the input sequence of the first round, performing splitting processing and position swapping processing; using the subsequence obtained after the splitting and position swapping processing in the previous round as the input sequence of this round, performing splitting processing and position swapping processing; iterating in sequence until the sequence length of the input sequence is 1, and outputting the position information of each element in the auxiliary sequence in the last round; wherein the splitting processing includes: splitting the input sequence from the middle to obtain two subsequences; the two subsequences include the front sequence segment and the rear sequence segment of the input sequence, and the length of the front sequence segment is the same as that of the rear sequence segment; the position swapping processing includes: extracting elements from the front sequence segment and the rear sequence segment in sequence, and comparing the elements extracted from the front sequence segment with the elements extracted from the rear sequence segment; when the elements extracted from the front sequence segment and the elements extracted from the rear sequence segment do not satisfy the specified sorting direction, swapping the positions of the elements extracted from the front sequence segment with the elements extracted from the rear sequence segment; Based on the position information of each element in the auxiliary sequence in the last round, extracting the sorted position information of each element in the first privacy sequence or the second privacy sequence held; Another business party that cooperates with the two business parties performs business collaborative processing according to the position information of the elements in the first privacy sequence and the second privacy sequence after being sorted.
2. The method according to claim 1, characterized in that Taking any round in the iterative steps as the current round, before performing the splitting process and the position exchange process on the input sequence of the current round, the method further includes: Calculate the ratio of the combined sequence length to the sequence length of the input sequence of the current round, and round up the ratio to obtain the sorted sequence value of the current round; Based on the order from front to back, extract the input sequence corresponding to the number of the sorted sequence values from the input sequence of the auxiliary sequence in the current round as the sorted input sequence of the current round; In the current round, a splitting process and a position exchange process are performed on the sorted input sequence.
3. The method according to claim 1, characterized in that Before constructing the initial auxiliary sequence, the method further includes: The first initial plaintext sequence or the second initial plaintext sequence held locally is sorted in a specified sorting direction and ciphertext processed to obtain a first privacy sequence or a second privacy sequence.
4. The method according to claim 1, characterized in that: The sequence length of the initial auxiliary sequence is greater than or equal to the first power of 2 of the combined sequence length.
5. The method according to claim 1, characterized in that The specified sorting direction is a sorting direction from small to large; accordingly, the reference element is an infinite element.
6. The method according to claim 1, characterized in that The specified sorting direction is a sorting direction from large to small; accordingly, the reference element is an infinitesimal parameter.
7. A business collaborative processing device based on privacy protection, characterized in that: Applicable to any one of two business parties holding sequences to be sorted in business collaborative processing, the device comprises: A sequence collaborative construction module, used to collaborate with the other business party of the two business parties to construct an initial auxiliary sequence based on the combined sequence length; wherein the sequence length of the initial auxiliary sequence is a value greater than or equal to the combined sequence length and a power of 2; the combined sequence length is the sum of the lengths of the first privacy sequence and the second privacy sequence respectively held by the two business parties; wherein the first privacy sequence and the second privacy sequence are sorted in a specified sorting direction; An element collaborative filling module is used to collaborate with the other business party of the two business parties to perform element filling processing on the initial auxiliary sequence to obtain an auxiliary sequence; wherein the element filling processing includes: filling each element of the first privacy sequence into the initial auxiliary sequence from front to back, filling each element of the second privacy sequence into the initial auxiliary sequence from back to front, and configuring a reference element corresponding to the specified sorting direction at a position without an element; The element position collaborative sorting module is used to collaborate with the other business party of the two business parties to perform the following iterative steps on the auxiliary sequence: using the auxiliary sequence as the input sequence of the first round, performing splitting processing and position swapping processing; then using the subsequence obtained after the previous round of splitting and position swapping processing as the input sequence of this round, performing splitting processing and position swapping processing; iterating in sequence until the sequence length of the input sequence is 1, and outputting the position information of each element in the auxiliary sequence in the last round; wherein the splitting processing includes: splitting the input sequence from the middle to obtain two subsequences; the two subsequences include the front sequence segment and the rear sequence segment of the input sequence, and the length of the front sequence segment is the same as that of the rear sequence segment; the position swapping processing includes: extracting elements from the front sequence segment and the rear sequence segment in sequence, and comparing the elements extracted from the front sequence segment with the elements extracted from the rear sequence segment; when the elements extracted from the front sequence segment and the elements extracted from the rear sequence segment do not satisfy the specified sorting direction, swapping the positions of the elements extracted from the front sequence segment with the elements extracted from the rear sequence segment; A position information extraction module, used to extract the sorted position information of each element in the first privacy sequence or the second privacy sequence based on the position information of each element in the auxiliary sequence in the last round; The business collaborative processing module is used to collaborate with another business party of the two business parties to perform business collaborative processing according to the position information of the elements in the first privacy sequence and the second privacy sequence after sorting.
8. The device according to claim 7, characterized in that The device also includes a sorting module and a ciphertext processing module; wherein, The sorting module is used to sort the first initial plaintext sequence or the second initial plaintext sequence held locally according to a specified sorting direction to obtain a first plaintext sequence or a second plaintext sequence; The ciphertext processing module is used to perform ciphertext processing on the first plaintext sequence or the second plaintext sequence held to obtain the first privacy sequence or the second privacy sequence.
9. A computer-readable storage medium having computer instructions stored thereon, characterized in that: When the instructions are executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.
10. A computer device comprising a memory, a processor and computer instructions stored in the memory, characterized in that: The processor executes the computer instructions to implement the steps of the method according to any one of claims 1 to 6.
