Patent Application
20160330016
The present invention relates to a technique of efficiently replicating an element included in a set.
Secure computation is a method to obtain a certain calculation result without restoring concealed values (see Non-patent Literature 1, for example). The method described in Non-patent Literature 1 performs concealing in which fragments of a value are distributed to three secure computation devices. The three secure computation devices perform calculations in cooperation, whereby the results of calculations such as addition and subtraction, constant addition, multiplication, constant multiplication, logical operation (NOT, logical AND, logical OR, exclusive OR), and data format conversion (integer, binary) can be maintained, without restoring a concealed value, in a state distributed to the three secure computation devices, that is, in a concealed state.
Non-patent Literature 2 recites a method of mapping these elements in accordance with a table by secure computation. In the method described in Non-patent Literature 2, a map in which an element (value) in a concealed table is replicated in the same sequence is realized in secure computation.
However, in a conventional technique, when the size of a column in which an element is replicated is N, O(N) times and O(log N) stages of multiplication has been necessary to replicate the element in the same column. This is true not only with a case where an element in a concealed table is replicated in the same column by secure computation but also with a case where an element in a set including ordered N elements is replicated as an element in the same set.
An object of the present invention is to reduce a calculation cost to replicate an element in a set including ordered elements.
To solve the above-described problem, processes described below are performed. A second set including a plurality of elements a5(1), . . . , a5(N) or a concealed text of the second set is obtained, where the second set is obtained by setting a replication source element a(f(h)) included in a first set to an element a(f(h))-a(f(h−1)) and setting elements other than the replication source in the first set to zero with respect to h=2, . . . , M. A third set or a concealed text of the third set is obtained using the second set or the concealed text of the second set, where the third set includes a first element b(1)=a5(1) and i=2, . . . , Nth element b(i)=b(i−1)+a5(i).
The first set includes a plurality of ordered elements a(1), . . . , a(N). A set of the elements a(1), . . . , a(N) includes a plurality of replication source elements a(f(1)), . . . , a(f(M)). An additive inverse of a replication source element a(f(h−1)) is −a(f(h−1)), where the order of replication source element a(f(h−1)) is before the replication source element a(f(h)) and is the closest to the replication source element a(f(h)). The first element in the second set is a5(1), and i=2, . . . , Nth element in the second set is a5(i).
According to the present invention, an element in a set can be replicated by addition, in which the calculation amount is small. This reduces calculation cost for replicating an element in a set including ordered elements.
Embodiments of the present invention will be described below.
[Outline]
Firstly, the outline of each embodiment will be described.
In each embodiment, processes described below are performed. A second set including a plurality of elements a5(1), . . . , a5(N) or a concealed text of the second set is obtained, where the second set is obtained by setting a replication source element a(f(h)) included in a first set to an element a(f(h))-a(f(h−1)) and setting elements other than the replication source in the first set to zero with respect to h=2, . . . , M. A third set or a concealed text of the third set is obtained using the second set or the concealed text of the second set. The third set includes a first element b(1)=a5(1) and i=2, . . . , Nth element b(i)=b(i−1)+a5(i).
The third set is a set in which the “replication source element” in the first set is replicated to all “elements other than the replication source” which are in the order after the “replication source element” and before the next replication source element in accordance with a certain order, and the “last replication source element” is replicated to all “elements other than the replication source” which are in the order after the “last replication source element.” By calculating the second set or the concealed text of the second set, where the second set is obtained by setting the replication source element a(f(h)) included in the first set to the element a(f(h))-a(f(h−1)) and setting the elements other than the replication source to zero, the above-described third set or the concealed text of the third set can be calculated by cumulative addition. The calculation amount of addition is very small compared with multiplication such as product used in a conventional technique. This reduces calculation cost for replicating an element in a set including ordered elements, compared with a conventional technique.
The “first set” indicates a set including a plurality of ordered elements a(1), . . . , a(N), where N is an integer equal to or larger than 2, for example, an integer equal to or larger than 3. The first set may be an N-dimensional vector consisting of N elements a(1), . . . , a(N), a set of N elements a(1), . . . , a(N) included in a N+A-dimensional vector where A is an integer equal to or larger than 1, a set of N elements a(1), . . . , a(N) included in a matrix having N or more elements, or other set. Among the elements a(1), . . . , a(N), an order is defined (that is, a plurality of elements a(1), . . . , a(N) are ordered). This order may be predetermined or determined each time processing is performed. The ordering method is not limited, any ordering is acceptable as long as each element corresponds to a different order. For example, when the first set is an N-dimensional vector whose 1 st, . . . , Nth elements are a(1), . . . , a(N) respectively, the element a(n) may corresponds to nth order where nε{1, . . . , N}, or the element a(n) may corresponds to N-n+1th order.
The set {a(1), . . . , a(N)} of the elements a(1), . . . , a(N) includes a plurality of replication source elements a(f(1)), . . . , a(f(M)), where, 2≦M≦N, for example, M<N. The “replication source elements” indicate elements serving as the replication sources, that is, elements to be replicated. Each element a(n) (where, nε{1, . . . , N}) is a replication source element a(f(m)) (where, mε{1, . . . , M}) or not a replication source element. In other words, each element a(n) holds {a(f(1)), . . . , a(f(M))}⊂{a(1), . . . , a(N)}, for example, {a(f(1)), . . . , a(f(M))}⊂{a(1), . . . , a(N)}. The replication source elements a(f(1)), . . . , a(f(M)) maintain the order relation in the elements a(1), . . . , a(N). That is, the orders in accordance with the orders of the elements a(1), . . . , a(N) correspond to the replication source elements a(f(1)), . . . , a(f(M)). For example, when N=10 and M=4 hold, the orders of the elements a(1), . . . , a(10) are respectively 1st, . . . , 10th, and the replication source elements are a(1), a(4), a(5), and a(9), the orders of the replication source elements a(1), a(4), a(5), and a(9) in a set {a(1), a(4), a(5), a(9)} are respectively 1st, 2nd, 3rd, and 4th. Furthermore, an element of which the order is before the replication source element a(f(h)) (where, hε{2, . . . , M}) and is the closest to the replication source element a(f(h)) is expressed as a(f(h−1)). In other words, the replication source element of which the order in a set formed only of replication source elements is immediately before the replication source element a(f(h)) is expressed as a(f(h−1)).
For each element, at least addition (additive operation) is defined, and the addition is expressed as “+”. Furthermore, for each element, at least an additive inverse is defined, and the additive inverse of an element ν is expressed as “−ν”. More specifically, the above a(f(h))-a(f(h−1)) represents the addition which adds the additive inverse −a(f(h−1)) to the element a(f(h)), and b(i−1)+a5(i) represents the addition which adds the element a5(i) to the element b(i−1). Examples of the addition include summation, exclusive OR, remainder, θ mod P, of a result of summation θ by P where P is an integer equal to or larger than 2. For example, when the addition is summation, a(f(h))-a(f(h−1)) is a subtracted value obtained by subtracting a(f(h−1)) from a(f(h)), and b(i−1)+a5(i) is an addition value obtained by adding the element a5(i) to the element b(i−1). Furthermore, an additive identity is expressed as “zero” or “0”. An additive generator is expressed as “1”.
A “concealed text of a” in the present embodiment indicates information obtained by concealing a with a method enabling secure computation. Concealing methods include publicly known secret sharing and encryption. Examples of “concealing methods enabling secure computation” include publicly known linear secret sharing and homomorphic encryption. With these methods, a binary operation ∘ is defined between α1 and α2, for example, and when α1∘α2=α3 is established, a concealed text E(α1) of α1 and a concealed text E(α2) of α2 can be used to calculate a concealed text E(α3)=E((α1)∘E(α2) of α3. In the same manner, when a unary operation ∘ is defined for α1 and ∘α1=a3 is established, for example, a concealed text E(α1) of α1 can be used to calculate a concealed text E(α3)=∘E(α1) of α3. These calculations neither require restoration process for the concealed texts E(α1) nor E(α2). Concealing, restoring, addition, subtraction, multiplication, and logical operation on secure computation can be executed by well-known techniques described in Non-patent Literatures 1 and 2, for example. Random permutation on secure computation can be executed by well-known techniques described in Non-patent Literature 2 or Reference Literature 1 (Sven Laur, Jan Willemson, and Bingsheng Zhang, “Round-efficient oblivious database manipulation”, In Xuejia Lai, Jianying Zhou, and Hui Li, editors, ISC, Vol. 7001 of LNCS, pp. 262-277, Springer, 2011), for example. A stable sort on secure computation can be executed by well-known techniques described in Non-patent Literature 2 or Reference Literature 2 (Koki Hamada, Dai Ikarashi, Koji Chida, Katsumi Takahashi, “A linear time sorting algorithm on secure function evaluation”, In SCIS, pp. 1-7, 2011), for example. These processing can be executed by other techniques such as one described in Reference Literature 3 (Koji Chida, Noburou Taniguchi, Go Yamamoto, Masato Okazaki, Osamu Shionoiri, Atsushi Kanai, “Implementation and Application of Secure Circuit Evaluation Based on the ElGamal Encryption”, CSS 2005, pp. 475-480, 2005), for example.
When a fourth set including the above element a(f(h))-a(f(h−1)) or a concealed text of the fourth set can be obtained in advance from a fifth set including the replication source elements a(f(1)), . . . , a(f(M)) or a concealed text of the fifth set, a first arithmetic unit may use this fourth set or the concealed text of the fourth set to obtain the second set or the concealed text of the second set (pre-calculation method). For example, when the fourth set or the concealed text of the fourth set can be obtained in advance, the first arithmetic unit may use this fourth set or the concealed text of the fourth set to obtain the second set or the concealed text of the second set, where the fourth set obtained by setting the replication source element a(f(h)) to a(f(h))-a(f(h−1)), the replication source element a(f(h)) included in a sixth set including the replication source elements a(f(1)), . . . , a(f(M)) ordered in accordance with the orders of the elements a(1), . . . , a(N). This further reduces the calculation cost. This method is suitable for a case where the first set or the concealed text of the first set can be obtained from the fifth set or the concealed text of the fifth set or the sixth set or the concealed text of the sixth set (hereinafter, “the fifth set and others”), that is, a case where “the fifth set and others” can be obtained before the first set or the concealed text of the first set (such as the method described in Non-patent Literature 2), for example.
Alternatively, the first arithmetic unit may use the first set or the concealed text of the first set to obtain the second set or the concealed text of the second set (post-calculation method). For example, the first arithmetic unit may use the first set or the concealed text of the first set to obtain a seventh set or a concealed text of the seventh set, where the seventh set is obtained by performing a stable sort of the elements a(1), . . . , a(N) included in the first set under a sorting criteria that is whether or not the element is a replication source, and the seventh set includes elements a(1), . . . , a(N) in which the orders of the elements a(f(1)), . . . , a(f(M)) are successive. The first arithmetic unit may further use the seventh set or the concealed text of the seventh set to obtain an eighth set or a concealed text of the eighth set, where eighth set is obtained by setting, with respect to h=2, . . . , M, a replication source element a(f(h)) included in the seventh set is set to a(f(h))-a(f(h−1)). The first arithmetic unit may then use the eighth set or the concealed text of the eighth set to obtain the second set or the concealed text of the second set, where the second set is obtained by sorting the elements in the eighth set and setting the elements other than the replication source to zero.
In this case, for example, the first arithmetic unit may use the concealed text of the first set to obtain the concealed text of the seventh set by secure computation. The first arithmetic unit may further obtain a concealed text of a ninth set, where the ninth set includes elements obtained by sorting elements p(a(1)), . . . , p(a(N)) representing the orders of the elements a(1), . . . , a(N) in accordance with the stable sort of the elements a(1), . . . , a(N). The first arithmetic unit may further use the concealed text of the seventh set to obtain the concealed text of the eighth set by secure computation. The first arithmetic unit may further use the concealed text of the eighth set and the concealed text of the ninth set to obtain, by secure computation, a concealed text of a tenth set and a concealed text of an eleventh set respectively, where the concealed text of the tenth set corresponding to the tenth set obtained by random permutation of the orders of the elements in the eighth set, the concealed text of the eleventh set corresponding to the eleventh set obtained by random permutation of the orders of the elements in the ninth set, while maintaining the correspondence relation between the elements in the eighth set and the elements in the ninth set. The tenth set is a set in which the orders of elements in the eighth set have been rearranged, and the eleventh set is a set in which the orders of elements in the ninth set have been rearranged. The first arithmetic unit may then use a restoring result of the concealed text of the eleventh set and the concealed text of the tenth set to obtain the concealed text of the second set which is a set obtained by sorting the elements in the tenth set and setting the elements other than the replication sources to zero.
Although the content of each element in the concealed text of the eighth set is concealed, the orders of the concealed texts of a(f(h))-a(f(h−1)) corresponding to the replication source elements are successive. Thus, if the ninth set is restored from the concealed text thereof and the elements representing the orders of the concealed texts of a(f(h))-a(f(h−1)) corresponding to the replication source elements are restored, information representing which element in the first set has been the replication source element is leaked to some extent. On the other hand, such information is not leaked if the elements representing the orders are obtained after random permutation on secure computation as described above.
When the “concealed text of the second set” and the “concealed text of the third set” are obtained in each embodiment, the first arithmetic unit obtains the concealed text of the second set by secure computation, and a second arithmetic unit uses the concealed text of the second set to obtain the concealed text of the third set by secure computation.
Each embodiment will be described below in detail with reference to the drawings.
In each embodiment, a concealed text of information a is denoted as [α]. For example, [α] having been concealed through secret sharing is a fragment (share) on the secret sharing and the concealed [α] through encryption is a cryptogram. The sth element of a vector β is denoted as β(s), and the entire vector obtained by concealing each element β(s) of the vector β is denoted as [β] and referred to as a concealed text [β] of the vector β.
An example will be described in each embodiment where the first set is an N-dimensional vector whose 1st, . . . , Nth elements are respectively the elements a(1), . . . , a(N), the second set is an N-dimensional vector a5 whose 1st, . . . , Nth element are respectively the elements a5(1), . . . , a5(N), and the third set is an N-dimensional vector b whose 1st, . . . , Nth elements are respectively elements b(1), . . . , b(N). The elements of these N-dimensional vectors are ordered as the 1st, . . . , Nth elements, respectively. The vectors a, a5, and b where N=10 and M=4 will be exemplified below.
where, w, x, y, and z which are respectively the 1st, 4th, 5th, and 9th element of the vector a are the replication source elements, and elements * other than these are the elements that are not the replication sources. “*” represents an arbitrary element. Although 2≦M<N in this example, the processing in the present embodiment can be performed even when 2≦M≦N.
In the present embodiment, a specific example of the above-mentioned pre-calculation method will be described.
<Configuration>
As illustrated in
<Pre-Calculation>
In the present embodiment, the vector a (the first set) is not input to the element replication device 11-k, but a concealed text [d1] of a vector d1 formed only of the replication source elements a(f(1)), . . . , a(f(M)) (the concealed text of the fifth set or the concealed text of the sixth set) is input to the element replication device 11-k, and stored in the storage 111-k. The respective elements a(f(1)), . . . , a(f(M)) of the vector d1 are ordered accordance with the orders of the elements a(1), . . . , a(N) of the vector a. In the example according to the present embodiment, the jth element d1(j) of the vector d1 is d1(j)=a(f(j)), where j ε{1, . . . , M}. The size N of the vector a and the orders of the elements a(1), . . . , a(N) are predetermined (in the present embodiment, a(1), . . . , a(N) are the 1st, . . . , Nth elements of the N-dimensional vector).
As illustrated in
<Element Replication>
Processing of element replication according to the present embodiment will be described with reference to
The cumulative addition arithmetic unit 115-k uses the concealed text [a5] of the vector a5 (the second set) to obtain, by secure computation, a concealed text [b] of the vector b (the third set) which is an N-dimensional vector whose 1st element is b(1)=a5(1) and i=2, . . . , Nth element is b(i)=b(i−1)+a5(i), and outputs the obtained concealed text [b](Step S115). The cumulative addition arithmetic unit 115-k, for example, firstly performs secure computation to obtain a concealed text [b(1)] of b(1)=a5(1) and then performs secure computation to obtain a concealed text [b(i)] of b(i)=b(i−1)+a5(i) with respect to each of i=2, . . . , N while increasing i by 1 from i=2, to obtain the concealed text [b] of the vector b.
For example, for the vector a exemplified in the expression (1), the relation between the vectors a5 and b described above is as below.
In the example described in the first embodiment, the concealed text [d1] of the vector d1 in which the orders are set in accordance with the orders of the elements a(1), . . . , a(N) of the vector a is input to the element replication device 11-k and stored in the storage 111-k. In the present modification, processing will be described in a case where a concealed text [d] of a vector d in which only the replication source elements a(f(1)), . . . , a(f(M)) are included but the orders are not set in accordance with the orders of the elements a(1), . . . , a(N) of the vector a is input to the element replication device 11-k and stored in the storage 111-k.
An element replication device 111′-k of a secure computation system 1′ according to the present modification consists of the element replication device 11-k in the first embodiment and a sorting unit 112′-k added thereto (
In the first embodiment and the first modification of the first embodiment, secure computation is applied to a concealed text of a vector serving as a processing object. However, with respect to a vector that may be disclosed, a plaintext vector may be a processing object instead of a concealed text of a vector, and calculation may be performed with respect to the plaintext vector, instead of secure computation. For example, in a case where a vector d may be disclosed, the sorting unit 112′-k sorts the vector d to obtain the vector d1. For example, when the vector d1 may be disclosed, the additive inverse arithmetic unit 113-k uses the vector d2 to calculate the vector d1, and the input information arithmetic unit 114-k may use the vector d2 as is. The additive inverse arithmetic unit 113-k uses the vector d2 to calculate the vector as, and the cumulative addition arithmetic unit 115-k may use the vector a5 to calculate the vector b, for example. In another case, a vector as a calculation result may be concealed and output.
In the present embodiment, a specific example of the above-mentioned post-calculation method will be described.
<Configuration>
As illustrated in
<Element Replication>
Processing of element replication according to the present embodiment will be described with reference to
It is assumed that a concealed text [α] of the vector a (the first set) and a concealed text [c] of a vector c indicating whether or not elements a(1), . . . , a(N) of the vector a are the replication source elements are input to the element replication device 21-k and stored in the storage 211-k. The nth element of the vector c is φ(a(n)). When a(n) is a replication source element, φ(a(n))=1, and when φ(a(n)) is not a replication source element, φ(a(n))=0. For example, for the vector a exemplified in the expression (1), the relation between the vectors a and c is as below.
Firstly, the order information generating unit 212-k (
The input information arithmetic unit 214-k extracts, from the storage 211-k, the concealed text [α] of the vector a (the first set), the concealed text [p] of the vector p, and the concealed text [c] of the vector c, obtains the concealed text [a5] of the vector a5 (the second set) by secure computation, and outputs the obtained concealed text [a5] (Step S214).
<<Details of Step S214>>
The details of Step S214 will be described with reference to
In the present embodiment, firstly, the stable sorting unit 2141-k extracts the concealed text [α] of the vector a and the concealed text [p] of the vector p from the storage 211-k. The stable sorting unit 2141-k uses the extracted concealed text [α] of the vector a and concealed text [p] of the vector p to obtain a concealed text [a1] of the vector a1 (the seventh set) and a concealed text [p1] of the vector p1 (the ninth set) by secure computation, and outputs the obtained concealed texts [a1] and [p1]. The vector a1 is obtained by performing a stable sort of the elements a(1), . . . , a(N) included in the vector a under a sorting criteria that is whether or not the elements are replication sources. The vector Pi is formed of the elements obtained by sorting the elements p(a(1)), . . . , p(a(N)) in accordance with the stable sort of these elements a(1), . . . , a(N) (Step S2141). The vector a1 is a vector in which only the elements a(1), . . . , a(N) are included and the orders of the replication source elements a(f(1)), . . . , a(f(M)) are successive. More specifically, the vector a1 is a vector in which only the elements a(1), . . . , a(N) are included, the ηth, . . . , η+M−1th elements of the vector a1 are the replication source elements a(f(1)), . . . , a(f(M)), and other elements are not the replication sources. η is an integer not less than 1 and not more than N+1-M. “the ηth element” indicates the element of which the order is ηth. For example, the vector a1 may be a vector in which the 1st, . . . , Mth elements of the vector a1 are the replication source elements a(f(1)), . . . , a(f(M)) and other elements are not the replication sources, or may be a vector in which the N-M+1th, . . . , Nth elements of the vector a1 are the replication source elements a(f(1)), . . . , a(f(M)) and other elements are not the replication sources. This stable sort is performed while the correspondence between the element a(n) of the vector a and the element p(a(n)) of the vector p is maintained. In other words, the correspondence between the element a(n) and the element p(a(n)) is maintained. For example, when the element a(n) becomes the n′th element of the vector a1 (where, n is an integer not less than 1 and not more than N) through this stable sort, the element p(a(n)) corresponding to the element a(n) is the n′th element of the vector p1. This is the order of the element in a plaintext, and the concealed element is not necessarily in this order. For example, when the vector a is as exemplified in the expression (1), p(a(n))=n−1 holds, the 1st, . . . , Mth elements of the vector a1 are the replication source elements a(f(1)), . . . , a(f(M)), and other elements are not the replication source elements, the relation between the vectors a1 and p1 described above is as below.
The additive inverse arithmetic unit 2142-k uses the concealed text [a1] of the vector a1 (the seventh set) to obtain, by secure computation, a concealed text [a2] of the vector a2 (the eighth set), and outputs the obtained concealed text [a2], where the vector a2 is an N-dimensional vector which is obtained by setting the elements at(1), a1(2), . . . , at(N) included in the vector a1 to elements a2(1)=a1(1), a2(2)=a1(2)-a1(1), . . . , a2(N)=a1(N)-a1(N−1) (Step S2142). For example, in the case of the vector a1 exemplified in the expression (2), the relation among the vectors p1, a1, and a2 is as below.
The sorting unit 2143-k uses the concealed text [a2] of the vector a2 (the eighth set) and the concealed text [p1] of the vector p1 (the ninth set) to obtain, by secure computation, a concealed text [a4] of the vector a4 which is an N-dimensional vector obtained by sorting the elements of the vector a2, and outputs the obtained concealed text [a4] (Step S2143). The vector a4 is a vector obtained by rearranging the orders of the elements of the vector a2 into the original orders corresponding to the orders of the elements of the vector a. More specifically, the vector a4 is a vector obtained by setting the element a(f(h))-a(f(h−1)) of the vector a2 to the order of the replication source element a(f(h)) of the vector a, setting the element a(f(1)) of the vector a2 to the order of the replication source element a(f(1)) of the vector a, and returning other elements of the vector a2 into the original orders of the vector a corresponding to the other elements.
<<Details of Step S2143>>
The details of Step S2143 will be described with reference to
The random permutation unit 2143a-k of the sorting unit 2143-k uses the concealed text [a2] of the vector a2 (the eighth set) and the concealed text [p1] of the vector p1 (the ninth set) to obtain, by secure computation, a concealed text [a3] of a vector a3 and a concealed text [p3] of a vector p3 respectively corresponding to an N-dimensional vector a3 (the tenth set) and an N-dimensional vector p3 (the eleventh set), and outputs the obtained concealed text [a3] and concealed text [P3], where the N-dimensional vector a3 is obtained through random permutation of the orders of elements a2(1), . . . , a2(N) of the vector a2, and the N-dimensional vector p3 is obtained through random permutation of the orders of elements p1(1), . . . , p1(N) of the vector p1, while maintaining the correspondence relation between an element a2(n) of the vector a2 and an element p1(n) of the vector p1 (where, nε{1, . . . , N}) (Step S2143a). The vector a3 is a vector obtained by rearranging the orders of the elements a2(1), . . . , a2(N) of the vector a2 and the vector p3 is a vector obtained by rearranging the orders of the elements p1(1), . . . , p1(N) of the vector p1. In other words, the random permutation unit 2143a-k uses the concealed text [a2] of the vector a2 (the eighth set) and the concealed text [p1] of the vector p1 (the ninth set) to obtain the concealed text [a3] of the vector a3 and the concealed text [p3] of the vector p3 by secure computation, and outputs the obtained concealed text [a3] and concealed text [p3]. The vector a3 is a vector obtained through random permutation of the orders of the elements a2(1), . . . , a2(N) of the vector a2 and the vector p3 is a vector obtained through random permutation of the orders of the elements p1(1), . . . , p1(N) of the vector p1. In this processing, the correspondence relation between the element a2(n) and the element p1(n) is maintained. For example, when the element a2(n) becomes the n′th element of the vector a3 through this random permutation, the element p1(n) is the n′th element of the vector p3. This is the order of the element in a plaintext, and the concealed element is not necessarily in this order. In the case of the vectors p1 and a2 exemplified in the expression (3), the relation between the vectors p3 and a3 is as below, for example.
The restoring unit 2143-b-k restores the vector p3 from the concealed text [p3] of the vector p3 and outputs the restored vector p3 (Step S2143b).
The order restoring unit 2143c-k uses the concealed text [a3] of the vector a3 to obtain, by secure computation, a concealed text [a4] of the vector a4 which is an N-dimensional vector obtained by sorting the elements of the vector a3 in accordance with the vector p3, and outputs the obtained concealed text [a4] (Step S2143c). The vector a4 is a vector obtained by rearranging the orders of the elements of the vector a3 into the original orders corresponding to the orders of the elements of the vector a. These orders can be determined by the elements of the vector p3 (refer to the expression (4)). More specifically, the vector a4 is the N-dimensional vector obtained by rearranging the elements of the vector a3 through the sort that rearranges the order of the elements of the vector p3 into the orders of the elements of the vector p. In the case of the vectors p3 and a3 exemplified in the expression (4), the vector a4 is as below, for example (this ends the description of <<Details of Step S2143>>).
The multiplication arithmetic unit 2144-k uses the concealed text [a4] of the vector a4 and the concealed text [c] of the vector c to obtain, by secure computation, the concealed text [a5] of the vector a5 (the second set), and outputs the obtained concealed text [a5], where an element a5(n) of the vector a5 is a result of multiplication of an element a4(n) of the vector a4 and the element c(n) of the vector c (for example, a multiplication result or remainder of a multiplication result) (where, nε{1, . . . , N}) (Step S2144). The element a4(n) is the nth element of the vector a4, the element c(n) is the nth element of the vector c, and the element a5(n) is the nth element of the vector as. This processing is processing for obtaining the concealed text [a5] of the vector a5 in which elements that are not the replication sources of the vector a4 are set to zero. For example, in the case of the vector c corresponding to the vector a exemplified in the expression (1) and the vector a4 exemplified in the expression (5), the relation among the vectors a4, c, and a5 is as below (this ends the description of <<Details of Step S214>>).
The concealed text [a5] of the vector a5 is input to the cumulative addition arithmetic unit 115-k, the processing described in the first embodiment is performed, and the concealed text [b] of the vector b is obtained. For example, in the case of the vector a5 in the expression (6), the relation between the vectors a5 and b is as below.
In the second embodiment, the sorting unit 2143-k performs the sort including random permutation by secure computation. Instead of this, the present modification performs a sort that does not include random permutation, by secure computation. Other matters are the same as those described in the second embodiment.
The input information arithmetic unit 214-k of an element replication device 21′-k of a secure computation system 2′ according to the present modification includes a sorting unit 2143′-k instead of the sorting unit 2143-k in the second embodiment (
In the second embodiment and the first modification of the second embodiment, secure computation is applied to a concealed text of a vector serving as a processing object. However, with respect to a vector that may be disclosed, a plaintext vector may be a processing object instead of a concealed text of a vector, and calculation may be performed with respect to the plaintext vector, instead of secure computation. For example, the order information generating unit 212-k may output the vector p, and the stable sorting unit 2141-k may perform a stable sort of the vectors a and p to obtain the vectors a1 and p1 and output the obtained vectors a1 and p1. The additive inverse arithmetic unit 2142-k may obtain the vector a2 from the vector a1 and output the obtained vector a2, and the sorting unit 2143-k or the sorting unit 2143′-k may use the vectors a2 and p1 to obtain the vector a4 and output the obtained vector a4. The multiplication arithmetic unit 2144-k may use the vectors a4 and c to obtain the vector a5 and output the obtained vector a5.
[Characteristics]
As described above, a replication source element is replaced by an addition result (for example, a subtraction result) that is obtained by adding an additive inverse of that replication source element to a replication source element immediately before that replication source element, whereby the replication processing can be implemented by an addition (for example, summation) which can perform many pieces of processing with a very light calculation. This can reduce the calculation cost compared with a conventional case. It should be noted that, in order to obtain the vector a5 or the concealed text [a5] from the vector a4 or the concealed text [a4], one phase of multiplication such as multiplication is necessary. However, the number of phases does not depend on N. Especially when preprocessing can be performed for a replication source element as in the first embodiment, only O(N) times of addition is required for the calculation for replication at the time of element replication.
It should be noted that the present invention is not limited to the embodiments described above. For example, various types of processing described above can be performed in time series in accordance with the descriptions as well as performed in parallel or individually in accordance with the capacity or the need of the device performing the processing. Needless to say, modifications also can be added as appropriate within the scope of the present invention.
When the configuration described above is implemented on a computer, the contents of processing function to be contained in each device is written by a program. With this program executed on the computer, the above-described processing function is executed on the computer. This program with the contents of processing written thereinto can be stored in a computer-readable recording medium. An example of the computer readable recording medium is a non-transitory recording medium. Examples of such a recording medium include a magnetic recording device, an optical disk, a magneto-optical recording medium, and a semiconductor memory.
Distribution of this program is implemented by sales, transfer, rental, and other transactions of a portable recording medium such as a DVD and CD-ROM with the program recorded thereon, for example. Furthermore, this program may be stored in a storage unit of a server computer and transferred from the server computer to other computers via a network so as to be distributed.
A computer that executes a program as described above first stores the program stored in a portable recording medium or transferred from a server computer in a storage unit of the computer, for example. When the processing is performed, the computer reads out the program stored in the recording medium of the computer and performs processing in accordance with the program thus read out. As another execution form of this program, the computer may directly read out the program from a portable recording medium and perform processing in accordance with the program. Furthermore, each time the program is transferred to the computer from the server computer, the computer may sequentially perform processing in accordance with the received program. Alternatively, what is called application service provider (ASP) type of services may be used to perform the processing described above, with which the program is not transferred from the server computer to the computer and the processing function is realized only with execution instructions and result acquisition.
In the present embodiments, the processing function of the present device is implemented with a predetermined program executed on a computer. However, at least part of the processing function may be implemented in a hardware manner.
The present invention can be used in the field of spreadsheet operation, for example, and especially, can be used in the field of software (encryption technology) for performing calculation that replicates an element as a part of a concealed spreadsheet in a concealed state, for example. The present invention can be used in element replication processing (such as spreadsheet processing) of a set used in the encryption technology, for example.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2014-006618 | Jan 2014 | JP | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/JP2014/081546 | 11/28/2014 | WO | 00 |
