Privacy-protected data sharing system and privacy-protected data sharing device

Abstract

To generate statistical information that excludes a correspondence relation with individuals without revealing contents of personal information to other devices.SOLUTION: In a data association system including a plurality of devices which maintain user data including a user ID and personal information about a user, each device includes: a de-identification processing unit which performs de-identification processing on the user data and generates de-identified data; an encryption unit which performs first encryption for encrypting the de-identified data using a self-encryption key held by a self-device and second encryption for encrypting the de-identified data subjected to the first encryption received from a partner device using a self-encryption key; and an aggregation processing unit which collates the de-identified data subjected to the second encryption with the de-identified data subjected to the second encryption received from the partner device and aggregates the number of de-identified data subjected to the second encryption whose user ID corresponding parts specified on the basis of structural information of the user data match each other as the number of target users.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] This disclosure relates to a privacy-protecting data linkage system and a privacy-protecting data linkage device. [Background technology]

[0002] When performing statistical analysis on user data managed on a database in information processing devices (devices that broadly include computers, servers, etc., hereinafter referred to as "devices") of multiple organizations, if the user data contains personal information, it is necessary to take some measures from the perspective of protecting privacy. As an example of the above measures, a technology is known that integrates both the data to be anonymized and the data that can be disclosed by having a system configuration with an interface that can handle both simultaneously, while keeping the data to be anonymized confidential (see Patent Document 1 below). In addition, a technology is known that protects personal information by applying both reversible and irreversible encryption to the ID of each piece of personal information when integrating user data containing personal information between multiple devices, making it difficult for a third party to recover the original ID (see Patent Document 2 below). [Prior art documents] [Patent Documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2011-081301 [Patent Document 2] Japanese Patent Publication No. 2010-211590 [Overview of the Initiative] [Problems that the invention aims to solve]

[0004] However, the technology in Patent Document 1 requires the selection of data to be anonymized from the input user data, and does not anticipate cases where the user data does not contain data that can be disclosed, i.e., cases where the entire input user data is subject to anonymization, so there is room for improvement in this respect. The technology in Patent Document 2 requires both reversible and irreversible encryption processing, and since the reversible encryption processing remains, there is room for improvement in that there is a possibility of recovery. In addition, both Patent Documents 1 and 2 do not completely eliminate the possibility that the correspondence between the processing result and an individual can be inferred, so there is room for improvement from the perspective of privacy protection.

[0005] This disclosure aims to generate statistical information that excludes any correspondence with individuals, without disclosing the content of personal information to other devices, by linking multiple devices that hold personal information. [Means for solving the problem]

[0006] The privacy protection data linkage system relating to this disclosure comprises a plurality of devices that hold user data including user IDs and personal information relating to users, each of which includes an anonymization processing unit that deidentifies the user data to generate anonymized data, an encryption unit that performs a first encryption using its own encryption key on the anonymized data, an encryption unit that performs a second encryption using its own encryption key on the first encrypted anonymized data received from the other party's device, and an aggregation processing unit that matches the second encrypted anonymized data with the second encrypted anonymized data received from the other party's device and aggregates the number of second encrypted anonymized data where the user ID corresponding portion identified based on predetermined structural information of the user data matches as the number of target users. [Effects of the Invention]

[0007] According to this disclosure, by linking multiple devices that hold personal information, it is possible to generate statistical information that excludes correspondences with individuals without disclosing the content of the personal information to other devices. [Brief explanation of the drawing]

[0008] [Figure 1] This is a diagram illustrating the configuration of a privacy-protecting data linkage system according to an embodiment of this disclosure. [Figure 2] This is a flowchart showing the processes performed in the privacy protection data linkage system according to the embodiment of this disclosure. [Figure 3] (a) is a diagram illustrating the deidentified data generation process using Company A's equipment, and (b) is a diagram illustrating the deidentified data generation process using Company B's equipment. [Figure 4] (a) is a diagram illustrating the grouping of deidentified data by Company A's equipment, and (b) is a diagram illustrating the grouping of deidentified data by Company B's equipment. [Figure 5] This diagram illustrates the encryption process of deidentified data using equipment from companies A and B. [Figure 6] This diagram provides supplementary explanation for the encryption process shown in Figure 5. [Figure 7] (a) is a diagram illustrating the aggregation process for the number of users who both used the service and visited the store, (b) is a diagram illustrating the aggregation process for the number of users who used the service but did not visit the store, (c) is a diagram illustrating the aggregation process for the number of users who did not use the service but visited the store, and (d) is a diagram illustrating the aggregation process for the number of users who neither used the service nor visited the store. [Figure 8] This is a diagram to explain the confidentiality process. [Figure 9] This is a diagram illustrating the configuration of the privacy-protecting data sharing system according to the first modified example. [Figure 10] This is a flowchart showing the processes performed in the privacy protection data linkage system according to the first modified example. [Figure 11]It is a configuration diagram of a privacy protection data sharing system according to a second modification example. [Figure 12] It is a flowchart showing processing executed in a privacy protection data sharing system according to a second modification example. [Figure 13] It is a diagram showing an example of the hardware configuration of each device.

Embodiments for Carrying Out the Invention

[0009] Hereinafter, embodiments of a privacy protection data sharing system and a privacy protection data sharing device according to the present disclosure will be described with reference to the drawings.

[0010] (Configuration of Privacy Protection Data Sharing System) As shown in FIG. 1, a privacy protection data sharing system 1 according to the present disclosure includes a plurality of devices 10 (corresponding to privacy protection data sharing devices) that can communicate with each other. As will be described later, statistical information regarding the number of target users is generated by the cooperation of one of the plurality of devices 10 and the partner device. That is, after one device and the partner device both execute "generation of non-identified data" and "encryption of non-identified data" described later, the one device that receives the double-encrypted non-identified data from the partner device executes "aggregation processing" and "concealment processing" to generate statistical information. In FIG. 1, for convenience of explanation, two devices 10A and 10B are shown, but three or more devices may be provided. In the following processing, an example in which the device 10A of Company A operates as "one device" and the device 10B of Company B operates as "partner device" will be described. However, each device 10 has the same functional block configuration described later and may operate as "one device" or as "partner device".

[0011] Furthermore, each device 10 stores user data, including a "User ID" and "Personal Information" about the user, in an internal memory not shown. Of these, the "User ID" is in a common data format across all devices, but the "Personal Information" is predetermined for each device and is not necessarily common across all devices. For example, the user data held by device 10A of Company A, shown in the upper part of Figure 3(a), includes "Name" and "Date of Birth," which correspond to the User ID, and "Gender," "Age Group," and "Usage Status," which correspond to personal information. On the other hand, the user data held by device 10B of Company B, shown in the upper part of Figure 3(b), includes "Name" and "Date of Birth," which correspond to the User ID, and "Visit Information" and "Purchase Information," which correspond to personal information. Thus, while the "Name" and "Date of Birth," which correspond to the User ID, are in a common data format, the types of information corresponding to personal information differ between the two. For example, the "usage status" included in the personal information of Company A above indicates whether or not the user is using the services provided by Company A (usage status); the "visit information" included in the personal information of Company B above indicates whether or not the user has visited a store operated by Company B (visit status); and the "purchase information" indicates the products purchased during the visit (personal computer (PC), microwave oven, etc.). The structural information of the user data held by each device 10 is shared between the devices in advance, and based on this structural information, the "user ID corresponding portion" in the double-encrypted anonymized data generated by each device 10 described later can be identified, and it is possible to match the "user ID corresponding portions" of the double-encrypted anonymized data described later.

[0012] The following describes the functional block configuration of each device 10, using device 10A of Company A as an example. As shown in Figure 1, device 10A comprises an anonymization processing unit 11, an encryption unit 12, an aggregation processing unit 13, and a confidentiality processing unit 14.

[0013] Of these, the de-identification processing unit 11 is a functional unit that targets user data held by its own device 10A, performs irreversible conversion processing to user IDs and processing for privacy protection of personal information, and generates de-identification data corresponding to user data where the personal information in the user data satisfies the conditions that must be met in order to extract the target user.

[0014] In this embodiment, device 10A of Company A and device 10B of Company B work together to create four different conditions for extracting target users, based on the combination of personal information "usage status" (used / not used) in Company A's user data and personal information "visit information" (visited / not visited) in Company B's user data. The process of aggregating the number of target users for each of these conditions and generating statistical information will be described later. In this embodiment, as an example of "irreversible transformation processing," an example of hashing processing using a hash function shared between devices 10 in advance will be described, but irreversible transformation processing is not limited to hashing processing. In addition, as an example of "privacy protection," k-anonymization will be described, but other methods such as l-diversity and t-approximation may be used, or a combination of two or more methods may be used.

[0015] The encryption unit 12 is a functional unit that generates double-encrypted anonymized data after double encryption using the self-encryption key held by its own device 10A and the other party's device 10B, by coordinating with the other party's device 10B as follows. Specifically, the encryption unit 12 performs a first encryption based on its own encryption key and keyed one-way commutative operations on the deidentified data generated by the deidentification processing unit 11 of its own device 10A, and exchanges the deidentified data after the first encryption with the other device 10B (i.e., the other device 10B that performed the first encryption based on the other device's encryption key and keyed one-way commutative operations on the deidentified data generated by the deidentification processing unit 11 of device 10B). Furthermore, it performs a second encryption based on its own encryption key and keyed one-way commutative operations on the deidentified data after the first encryption received from the other device 10B to generate double-encrypted deidentified data after double encryption using its own encryption key and the other device's encryption key. At this time, the encryption unit 12 of the other device 10B also performs a second encryption based on the other device's encryption key and keyed one-way commutative operations in the same manner as above to generate double-encrypted deidentified data after double encryption using its own encryption key and the other device's encryption key. Furthermore, the encryption process performed by the encryption unit 12 on the anonymized data is carried out separately for the "part corresponding to the user ID" and the "part corresponding to personal information" in the anonymized data.

[0016] In the encryption described above, the de-identified data after the first encryption, exchanged between devices 10A and 10B, is encrypted based on the other party's encryption key, which is not revealed to its own device. Therefore, the device cannot decrypt the de-identified data and cannot recognize the contents of the de-identified data (personal information, etc.). Similarly, the other party's device cannot decrypt the de-identified data after the first encryption obtained through the exchange and cannot recognize the contents of the de-identified data (personal information, etc.). In this way, the encryption unit 12 can generate double-encrypted de-identified data after double encryption using its own encryption key and the other party's encryption key, without revealing either the personal information in the unencrypted user data held by its own device 10A or its own encryption key to the other party's device 10B. Furthermore, since both encryption processes are based on keyed, one-way, commutative operations, it is possible to determine whether the two double-encrypted, anonymized data generated in one's own device 10A and the double-encrypted, anonymized data generated in the other party's device 10B correspond to the user data of the same user based on the matching or mismatch of the bit sequences of the user ID corresponding portion, which are identified based on the structural information of the user data shared in advance between devices 10A and 10B.

[0017] The aggregation processing unit 13 is a functional unit that aggregates the number of target users by comparing the double-encrypted anonymized data generated by the encryption unit 12 of the counterparty's device 10B with the double-encrypted anonymized data generated by the encryption unit 12 of its own device 10A, and counting the number of double-encrypted anonymized data where the "part corresponding to the user ID" matches. As described above, based on the matching or mismatch of the "part corresponding to the user ID" identified based on the data structure of the user data between these two double-encrypted anonymized data, it is possible to determine whether the two double-encrypted anonymized data correspond to the user data of the same user. Therefore, the aggregation processing unit 13 aggregates the "number of target users" which corresponds to the number of double-encrypted anonymized data where the personal information satisfies the conditions for a target user by counting the number of double-encrypted anonymized data where the "part corresponding to the user ID" matches as a result of the above matching, that is, the number of double-encrypted anonymized data that are determined to correspond to the user data of the same user.

[0018] The concealment processing unit 14 is a functional unit that performs concealment processing on the aggregated results obtained by the aggregation processing unit 13 to generate concealed statistical information. In this embodiment, as an example of "concealment processing," an example of concealment processing that adds noise of a certain intensity or higher, that is, concealment processing based on differential privacy, is described, but concealment processing is not limited to concealment processing based on differential privacy.

[0019] Each device 10 constituting the privacy protection data linkage system 1 is equipped with the functional units described above, but the details of these functions will be clarified in the following explanation of the processes based on Figures 2 to 8.

[0020] (Processing performed in the privacy-protected data sharing system) The following describes an example in which device 10A of Company A operates as "device 1" and device 10B of Company B operates as "the other party's device".

[0021] In device 10A of Company A, the de-identification processing unit 11 k-anonymizes the "target attribute information," which is personal information in the user data held by its device 10A (step A1), performs a hashing process on the "user ID" in the user data using a hash function that has been shared among the devices 10 in advance, and then discards the salt used in the hashing process (step A2). Supplementing the processes of steps A1 and A2 above with reference to Figure 3(a), in an example where the user data to be processed includes the name "Taro Yamada" and date of birth "January 1, 2000" corresponding to the "user ID," for subsequent processing, the name and date of birth are changed to the format "Taro Yamada 20000101" in the format "name + YYYYMMDD," and the gender, age, and usage status corresponding to the "target attribute information" are k-anonymized. Furthermore, a hashing process is performed on the "user ID" using the salt to convert the user ID into a hash value, and then the salt is discarded. This results in the acquisition of anonymized data containing the hash value "a12f7d9" corresponding to the "User ID" and k-anonymized "Target Attribute Information".

[0022] As mentioned above, in this embodiment, we will explain an example in which a total of four conditions corresponding to combinations of personal information "usage status" (used / not used) in Company A's user data and personal information "visit information" (visited / not visited) in Company B's user data are used as conditions for extracting target users, and the number of target users is aggregated for each of these conditions to generate statistical information. In this case, the de-identification processing unit 11 of Company A's device 10A binarizes the target attribute information (step A3) and separates it into bit sequences of the target attribute information (step A4). Here, the processing in steps A3 and A4 above will be explained in more detail based on Figure 4(a). Note that in Figure 4(a), for the sake of ease of explanation, an example is shown in which only "usage status" (used / not used) is considered as target attribute information corresponding to personal information. Similarly, Figure 4(b), which will be described later, also shows an example in which only "visit information" (visited / not visited) is considered as target attribute information corresponding to personal information.

[0023] As shown in the upper part of Figure 4(a), if the deidentified data to be processed includes a "hash value" corresponding to the user ID and a "usage status (1: present, 0: absent)" corresponding to the target attribute information, the deidentification processing unit 11 binarizes the target attribute information according to the "usage status (1: present, 0: absent)". At this time, as shown in the middle part of Figure 4(a), for example, the data with the hash value "4a3" is used, so the "usage present" bit is converted to "1" and the "usage absent" bit is converted to "0", while the data with the hash value "3c0" is not used, so the "usage present" bit is converted to "0" and the "usage absent" bit is converted to "1". The deidentification processing unit 11 then separates the deidentified data according to the bit sequence of the target attribute information (here, "10" or "01"), and as a result, as shown in the lower part of Figure 4(a), it is separated into a group where the bit sequence of the target attribute information is "10" and a group where the bit sequence of the target attribute information is "01".

[0024] Similarly, in device 10B of company B, the de-identification processing unit 11 k-anonymizes the "target attribute information" in the user data held by device 10B (step B1), hashes the "user ID" in the user data using a hash function that has been shared among the devices 10 in advance, and then discards the salt used in the hashing (step B2). To elaborate based on Figure 3(b), in an example where the raw data to be processed includes the name "Yamada Taro" and date of birth "January 1, 2000" corresponding to the "User ID", the name and date of birth are formatted to "Yamada Taro 20000101" in the format "Name + YYYYMMDD" (i.e., YYYY represents the year, MM represents the month, and DD represents the day) for subsequent processing. At the same time, the visited store and purchase information corresponding to the "Target Attribute Information" are k-anonymized. Furthermore, hashing is performed on the "User ID" using a salt to convert the User ID into a hash value, and then the salt is discarded. As a result, anonymized data is obtained that includes the hash value "a12f7d9" corresponding to the "User ID" and the k-anonymized "Target Attribute Information".

[0025] Furthermore, the deidentification processing unit 11 of Company B's device 10B binarizes the target attribute information (step B3) and separates it into bit sequences of the target attribute information (step B4). To further explain the processing in steps B3 and B4 based on Figure 4(b), in the example where the deidentification data to be processed includes a "hash value" corresponding to the user ID and "visit information (1: present, 0: absent)" corresponding to the target attribute information, as shown in the upper part of Figure 4(b), the deidentification processing unit 11 binarizes the target attribute information according to the "visit information (1: present, 0: absent)". At this time, as shown in the middle part of Figure 4(b), for example, the data with the hash value "4a3" is converted to data where the "visit information present" bit is "1" and the "visit information absent" bit is "0" because there is visit information, and the data with the hash value "3c0" is converted to data where the "visit information present" bit is "0" and the "visit information absent" bit is "1" because there is no visit information. The deidentification processing unit 11 then separates the deidentified data for each bit sequence of the target attribute information (here, "10" or "01"), and as a result, as shown in the lower part of Figure 4(b), the data is separated into a group where the bit sequence of the target attribute information is "10" and a group where the bit sequence of the target attribute information is "01".

[0026] Next, device 10A from company A and device 10B from company B work together to perform encryption of the anonymized data as shown in Figure 5.

[0027] In device 10A of company A, the encryption unit 12 performs the first encryption on the deidentified data generated by the deidentification processing unit 11 of device 10A, based on the encryption key held by device 10A and keyed one-way commutative operations (step A5 in Figure 2; process (1A) in Figure 5). Similarly, in device 10B of company B, the encryption unit 12 performs the first encryption on the deidentified data generated by the deidentification processing unit 11 of device 10B, based on the encryption key held by device 10B and keyed one-way commutative operations (step B5 in Figure 2; process (1B) in Figure 5). Then, the deidentified data after the first encryption is exchanged between devices 10A and 10B (steps A6, B6 in Figure 2; process (2) in Figure 5). The de-identified data exchanged between devices 10A and 10B after the first encryption is encrypted using the other party's encryption key, which is not revealed to the device itself. Therefore, the device itself cannot decrypt the de-identified data and cannot recognize its contents (personal information, etc.). Similarly, the other party's device cannot decrypt the de-identified data obtained through the exchange after the first encryption and cannot recognize its contents (personal information, etc.).

[0028] Then, in device 10A of company A, the encryption unit 12 performs a second encryption based on the encryption key held by device 10A and keyed one-way commutative operations on the deidentified data after the first encryption received from the other party's device 10B, thereby generating double-encrypted deidentified data after double encryption using the encryption key of device 10A and the encryption key of device 10B (Step A7 in Figure 2; Processing (3A) in Figure 5). Similarly, in device 10B of company B, the encryption unit 12 performs a second encryption based on the encryption key held by device 10B and keyed one-way commutative operations on the deidentified data after the first encryption received from the other party's device 10A, thereby generating double-encrypted deidentified data after double encryption using the encryption key of device 10A and the encryption key of device 10B (Step B7 in Figure 2; Processing (3B) in Figure 5).

[0029] Furthermore, the encryption unit 12 of Company B's device 10B transmits the double-encrypted non-identified data obtained in step B7 to the encryption unit 12 of Company A's device 10A (step B8; process (4) in FIG. 5), and the encryption unit 12 of Company A's device 10A receives the double-encrypted non-identified data from Company B's device 10B (step A8).

[0030] Here, using FIG. 6, the above encryption process will be supplemented and explained. Let the data of Company A be a, the data of Company B be b, the secret key held by Company A's device 10A be ka, the secret key held by Company B's device 10B be kb, and the process of encrypting x with the secret key k be f ka , (x) (where f k (x) is a keyed one-way convertible hash function), then by the first encryption, f ka (a) and f kb (b) are obtained, and these are exchanged between devices 10A and 10B. Next, in Company A's device 10A, the second encryption using Company A's secret key ka is executed to obtain f ka (f kb (b)), and in Company B's device 10B, the second encryption using Company B's secret key kb is executed to obtain f kb (f ka (a)). Then, f kb (f ka (a)) is transmitted from Company B's device 10B to Company A's device 10A.

[0031] As described above, both Company A's device 10A and Company B's device 10B execute encryption twice without revealing the secret keys ka and kb held by their own devices to the other party's device. Also, since f k (x) is a keyed one-way convertible hash function, from the characteristics of the convertible hash function, when a = b, f kb (f ka (a)) = f ka (f kb (b)) holds.

[0032] Next, in device 10A of Company A, the aggregation process for each combination of conditions is performed as shown in Figure 7. Specifically, the aggregation processing unit 13 of device 10A of Company A matches the double-encrypted, deidentified data after the second encryption for each combination of conditions (step A9 in Figure 2), and aggregates the number of target users by counting the number of data where the "user ID portion," which is identified based on the structural information of the user data shared in advance between devices 10A and 10B, matches (step A10 in Figure 2).

[0033] Here, we will use an example of matching double-encrypted, deidentified data obtained in the processes shown in Figures 4(a) and 4(b) to provide a supplementary explanation of the processes in steps A9 and A10 described above.

[0034] For the combination of conditions "used and visited" shown in Figure 7(a), only the data with a hash value of 4a3 was obtained as data where the user ID (hash value) portion after double encryption matched (matching was possible), so the count value of the number of data where the user ID portion after double encryption matched is "1" as the aggregated result. For the combination of conditions "used and not visited" shown in Figure 7(b), only the data with a hash value of 6f8 was obtained as data where the user ID portion after double encryption matched (matching was possible), so the count value of the number of data where the user ID portion after double encryption matched is "1" as the aggregated result.

[0035] Similarly, for the combination of conditions "no usage and store visit" shown in Figure 7(c), two data points were obtained (matchable) as data with matching user ID portions after double encryption: data with hash value 1e5 and data with hash value 5d9. Therefore, the count value of the number of data points with matching user ID portions after double encryption is "2". For the combination of conditions "no usage and no store visit" shown in Figure 7(d), only data with a matching user ID portion after double encryption was obtained (matchable) as data with matching user ID portions: data with hash value 3c0. Therefore, the count value of the number of data points with matching user ID portions after double encryption is "1".

[0036] Next, in device 10A of Company A, the concealment processing unit 14 performs concealment processing based on differential privacy on the aggregated results obtained in step A10, and generates and outputs statistical information from which correspondence with individuals has been excluded (step A11 in Figure 2). To explain using the aggregated results obtained in Figures 7(a) to 7(d) above as an example, the concealment processing unit 14 performs concealment processing based on differential privacy on the aggregated results for each of the four conditions, which are combinations of "usage present / not used" and "visit present / not visited," as shown in Figure 8, and generates and outputs statistical information for each of the four conditions from which correspondence with individuals has been excluded (n1, n2, n3, n4 in the example in Figure 8).

[0037] As described above, by coordinating between device 10A of Company A and device 10B of Company B, statistical information that excludes personal information is generated without disclosing personal information in the unencrypted user data held by each device, nor the encryption keys of each device, to the other device. In the example above, statistical information can be generated for each of the four conditions based on the combination of "usage / non-use" of Company A's service and "visit / non-visit" of Company B's store. This enables data linkage between Company A and Company B while protecting privacy, and allows for the acquisition of useful statistical information. Such useful statistical information can be effectively utilized by both Company A and Company B for advertising and customer acquisition strategies.

[0038] Furthermore, if the conditions that personal information must satisfy for the extraction of target users include multiple conditions, the de-identification processing unit 11 generates de-identified data (grouped de-identified data) corresponding to groups of user data where the personal information satisfies each of the multiple conditions. Therefore, statistical information regarding the number of target users that satisfy each of the multiple conditions can be obtained.

[0039] Furthermore, the encryption unit 12 works in conjunction with the encryption unit 12 of the other party's device to perform the series of processes shown in Figures 5 and 6 (first encryption → data exchange → second encryption). This allows both Company A's device 10A and Company B's device 10B to perform a total of two encryptions without revealing their own secret keys ka and kb to the other party's device. As a result, they can generate double-encrypted anonymized data after double encryption using the secret keys ka and kb.

[0040] Furthermore, the irreversible transformation process performed by the deidentification processing unit 11 includes a hashing process. Therefore, it eliminates the need to apply a special irreversible transformation process and allows the application of a general irreversible transformation process such as hashing, thereby improving ease of implementation.

[0041] Furthermore, the deidentification processing unit 11 discards the salt used in the hashing process after performing the hashing process on the user ID. By discarding the salt used in the hashing process in this way, it is possible to ensure that the possibility of data recovery is completely eliminated.

[0042] Furthermore, the privacy protection performed by the de-identification processing unit 11 includes at least one of k-anonymization, l-diversity, and t-approximation. Therefore, it eliminates the need to apply special privacy protections and allows the application of general privacy protections such as k-anonymization, l-diversity, and t-approximation, thereby improving ease of implementation.

[0043] Furthermore, the concealment processing performed by the concealment processing unit 14 includes concealment processing based on differential privacy. Therefore, it eliminates the need to apply special concealment processing and allows the application of general concealment processing based on differential privacy, thereby improving ease of implementation.

[0044] (Regarding variations) The privacy protection data linkage system described above in this disclosure may adopt the following first and second variations.

[0045] The first modification involves generating statistical information regarding the number of target users through the cooperation of three devices: "device one," "the other party's device," and a "third device," which are part of the multiple devices that make up the privacy-protecting data sharing system. The second modification involves the privacy-protecting data sharing system being composed of two types of devices: multiple encryption devices that perform encryption processing, and aggregation devices that perform aggregation processing and beyond. Each modification will be explained in turn below.

[0046] (First variation) As shown in Figure 9, the privacy protection data linkage system 1A according to the first modified example is a system that generates statistical information regarding the number of target users through the linkage between a total of three devices: "device one," "the other party's device," and "a third device." Figure 9 shows an example in which the privacy protection data linkage system 1A is configured to include multiple devices 10A, 10B, and 10C. For the sake of explanation, Figure 9 shows three devices 10A to 10C, but it may include four or more devices.

[0047] In the first modified example, we describe an example in which device 10A of Company A operates as "device one," device 10B of Company B operates as "the other party's device," and device 10C of Company C operates as "the third device." Each device has the same functional block configuration and can operate as "device one," as "the other party's device," or as "the third device."

[0048] The functions and effects of the "de-identification processing unit 11," "encryption unit 12," "aggregation processing unit 13," and "confidentiality processing unit 14" provided by each device are the same as those described in the embodiments of this disclosure above, so redundant explanations are omitted here. The structural information of the user data held by each device is shared between the devices in advance, and based on this structural information, the "user ID corresponding portion" in the double-encrypted de-identified data generated in devices 10A and 10B, respectively, can be identified, and the "user ID corresponding portions" of the double-encrypted de-identified data described later can be matched.

[0049] Figure 10 shows a series of processes performed in the privacy protection data linkage system 1A of the first modified example. The same reference numerals are used for processes that are the same as those described in Figure 2 in the embodiments of this disclosure described above. In Figure 10, the processes related to the generation of anonymized data (steps A1 to A4) performed by the anonymization processing unit 11 of Company A's device 10A, the processes related to the encryption of anonymized data (steps A5 to A8S) performed by the encryption unit 12 of Company A's device 10A, the processes related to the generation of anonymized data (steps B1 to B4) performed by the anonymization processing unit 11 of Company B's device 10B, and the processes related to the encryption of anonymized data (steps B5 to B8) performed by the encryption unit 12 of Company B's device 10B are almost the same as the processes in Figure 2 described above, so redundant explanations are omitted. However, in steps A8S and B8, the double-encrypted anonymized data after the second encryption is transmitted to Company C's device 10C, which operates as a "third device".

[0050] The double-encrypted, anonymized data from device 10A of Company A and device 10B of Company B are received by the aggregation processing unit 13 of device 10C of Company C (step C1). Subsequently, the aggregation processing unit 13 of device 10C of Company C, similar to step A9 in Figure 2 described above, refers to the structural information of user data that has been shared between the devices in advance, and matches the double-encrypted, anonymized data from device 10A with the double-encrypted, anonymized data from device 10B (step C2). Similar to step A10 in Figure 2 described above, the number of data where the user ID portion matches is counted to aggregate the number of target users (step C3). Furthermore, the concealment processing unit 14 of device 10C of Company C, similar to step A11 in Figure 2 described above, performs concealment processing based on differential privacy on the aggregation results obtained in step C3, generates statistical information that excludes correspondences with individuals, and outputs it to device 10A of Company A and device 10B of Company B (step C4).

[0051] Even in the first modified example described above, by enabling mutual data exchange between Company A's device 10A, Company B's device 10B, and Company C's device 10C while protecting privacy, it is possible to generate statistical information that excludes correspondences with individuals without disclosing personal information or encryption keys within the unencrypted user data held by Companies A and B to other devices. In this way, data exchange between Companies A, B, and C can be achieved while protecting privacy, and useful statistical information can be obtained. Such useful statistical information can be effectively utilized by both Companies A and B, which hold the user data, for advertising and customer acquisition strategies.

[0052] (Second variation) As shown in Figure 11, the privacy protection data linkage system 1B according to the second modified example consists of two types of devices: encryption devices 10X and 10Y that perform encryption processing, and aggregation device 10Z that performs aggregation processing and beyond. The system generates statistical information regarding the number of target users through the cooperation of these three devices: encryption devices 10X and 10Y, and aggregation device 10Z. For the sake of explanation, Figure 11 shows two encryption devices 10X and 10Y and one aggregation device 10Z, but it may be equipped with three or more encryption devices and two or more aggregation devices.

[0053] The second modification is similar to the first modification described above in that it generates statistical information regarding the number of target users through the cooperation of a total of three devices, but it differs in that it has two types of devices: an encryption device that specializes in performing encryption processing, and an aggregation device that specializes in performing aggregation processing and beyond.

[0054] Therefore, as shown in Figure 11, each of the encryption devices 10X and 10Y is equipped with the "de-identification processing unit 11" and the "encryption unit 12" from the components of the devices in Figures 1 and 9 described above, and the aggregation device 10Z is equipped with the "aggregation processing unit 13" and the "confidentiality processing unit 14" from the components of the devices in Figures 1 and 9 described above. The functions and effects of each of these components are the same as those described in the embodiments of this disclosure described above, so redundant explanations are omitted here.

[0055] Figure 12 shows a series of processes performed in the second modified example of the privacy-protected data linkage system 1B. Processes similar to those described in Figure 10 in the first modified example are denoted by the same reference numerals. As is clear from comparing Figure 12 with Figure 10, the series of steps A1-A8S, B1-B8, and C1-C4 are the same as the processes in Figure 10.

[0056] Therefore, the deidentification processing unit 11 and encryption unit 12 of the encryption device 10X in Figure 11 perform the same operations as the deidentification processing unit 11 and encryption unit 12 of Company A's device 10A in Figure 9, respectively, and the deidentification processing unit 11 and encryption unit 12 of the encryption device 10Y in Figure 11 perform the same operations as the deidentification processing unit 11 and encryption unit 12 of Company B's device 10A in Figure 9, respectively. At this time, the double-encrypted deidentified data generated by a total of two encryptions while protecting privacy in each of the encryption devices 10X and 10Y is received by the aggregation processing unit 13 of the aggregation device 10Z (step C1 in Figure 12), and the aggregation processing unit 13 and concealment processing unit 14 of the aggregation device 10Z perform the same operations as the aggregation processing unit 13 and concealment processing unit 14 of Company C's device 10C in Figure 9, respectively. As a result, the aggregated results obtained in step C3 of Figure 12 are subjected to anonymization processing based on differential privacy, generating statistical information from which correspondences with individuals have been excluded, and output to encryption devices 10X and 10Y, respectively.

[0057] In the second modified configuration described above, data exchange between the encryption devices 10X and 10Y and the aggregation device 10Z is achieved while protecting privacy. This allows for the generation of statistical information that excludes personal information and encryption keys within the unencrypted user data held by the encryption devices 10X and 10Y, without disclosing them to other devices. In this way, data exchange between the encryption devices 10X and 10Y and the aggregation device 10Z is achieved while protecting privacy, enabling the acquisition of useful statistical information. This useful statistical information can be effectively utilized by both the organization holding the encryption device 10X and the organization holding the encryption device 10Y for advertising and customer acquisition strategies.

[0058] (Explanation of terms, explanation of hardware configuration (Figure 13), etc.) The block diagrams used in the above descriptions of embodiments and modifications show functional units. These functional blocks (components) are realized by any combination of at least one of hardware and software. Furthermore, the method of realizing each functional block is not particularly limited. That is, each functional block may be realized using one device that is physically or logically coupled, or it may be realized using two or more physically or logically separated devices that are directly or indirectly connected (for example, using wired or wireless connections). A functional block may also be realized by combining software with the one or more devices described above.

[0059] Functions include, but are not limited to, judgment, decision, judgment, calculation, calculation, processing, derivation, investigation, exploration, confirmation, reception, transmission, output, access, resolution, selection, selection, establishment, comparison, assumption, expectation, assumption, broadcasting, notifying, communicating, forwarding, configuring, reconfiguring, allocating (mapping), and assigning. For example, a functional block (configuration part) that enables transmission is called a transmitting unit or transmitter. As mentioned above, the method of implementation is not particularly limited.

[0060] For example, the apparatus in one embodiment of the present disclosure may function as a computer that performs the processing in this embodiment. Figure 13 is a diagram showing an example of the hardware configuration of apparatus 10 according to one embodiment of the present disclosure. Apparatus 10 is a collective term for the aforementioned apparatuses 10A to 10C, encryption devices 10X and 10Y, and aggregation device 10Z. Physically, apparatus 10 may be configured as a computer device including a processor 1001, memory 1002, storage 1003, communication device 1004, input device 1005, output device 1006, bus 1007, etc.

[0061] In the following explanation, the term "device" can be replaced with "circuit," "device," "unit," etc. The hardware configuration of device 10 may include one or more of the devices shown in the figure, or it may be configured to omit some of the devices.

[0062] Each function in the device 10 is realized by loading predetermined software (programs) onto hardware such as the processor 1001 and memory 1002, which allows the processor 1001 to perform calculations, control communication by the communication device 1004, and control at least one of data reading and writing in the memory 1002 and storage 1003.

[0063] The processor 1001 controls the entire computer, for example, by running an operating system. The processor 1001 may consist of a central processing unit (CPU) that includes interfaces with peripheral devices, control units, arithmetic units, registers, and so on.

[0064] Furthermore, the processor 1001 reads programs (program code), software modules, data, etc., from at least one of the storage 1003 and the communication device 1004 into the memory 1002 and executes various processes accordingly. The program used is one that causes the computer to execute at least a part of the operations described in the above embodiment. Although the above processes have been described as being executed by one processor 1001, they may be executed simultaneously or sequentially by two or more processors 1001. The processor 1001 may be implemented by one or more chips. The program may also be transmitted from a network via a telecommunications line.

[0065] Memory 1002 is a computer-readable recording medium and may consist of at least one of the following: ROM (Read Only Memory), EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable ROM), RAM (Random Access Memory), etc. Memory 1002 may also be called a register, cache, main memory, etc. Memory 1002 can store executable programs (program code), software modules, etc., for carrying out a wireless communication method according to one embodiment of the present disclosure.

[0066] Storage 1003 is a computer-readable recording medium and may consist of at least one of the following: an optical disc such as a CD-ROM (Compact Disc ROM), a hard disk drive, a flexible disk, a magneto-optical disk (e.g., a compact disc, a digital multipurpose disc, a Blu-ray® disc), a smart card, flash memory (e.g., a card, a stick, a key drive), a floppy® disk, a magnetic strip, etc. Storage 1003 may also be called an auxiliary storage device. The above-mentioned storage medium may be, for example, a database including at least one of memory 1002 and storage 1003, or other suitable medium.

[0067] The communication device 1004 is hardware (transceiver / receiver device) for communicating between computers via at least one of a wired network and a wireless network, and is also referred to as a network device, network controller, network card, communication module, etc.

[0068] The input device 1005 is an input device that accepts input from the outside (e.g., a keyboard, mouse, microphone, switch, button, sensor, etc.). The output device 1006 is an output device that outputs to the outside (e.g., a display, speaker, LED lamp, etc.). The input device 1005 and the output device 1006 may be configured as an integrated unit (e.g., a touch panel). Furthermore, each device such as the processor 1001 and memory 1002 are connected by a bus 1007 for communicating information. The bus 1007 may be configured using a single bus, or different buses may be configured for each device.

[0069] Each aspect / embodiment described herein may be used individually, in combination, or switched between as needed during implementation. Furthermore, notification of specific information (e.g., notification that "X is") is not limited to explicit notification, but may also be implicit (e.g., by not providing such notification).

[0070] Although the present disclosure has been described in detail above, it will be clear to those skilled in the art that the present disclosure is not limited to the embodiments described herein. The present disclosure can be implemented in modified and altered forms without departing from the intent and scope of the present disclosure as defined by the claims. Therefore, the descriptions in the present disclosure are illustrative and not intended to be restrictive in any way.

[0071] The processing procedures, sequences, flowcharts, etc., of each aspect / embodiment described herein may be reordered, provided they are consistent. For example, the methods described herein present various step elements in an exemplary order and are not limited to that specific order.

[0072] Input and output information may be stored in a specific location (e.g., memory) or managed using a management table. Input and output information may be overwritten, updated, or appended to. Output information may be deleted. Input information may be transmitted to other devices.

[0073] In this disclosure, the phrase "based on" does not mean "based solely on" unless otherwise specified. In other words, the phrase "based on" means both "based solely on" and "based at least on."

[0074] Where the terms “include,” “including,” and variations thereof are used in this disclosure, these terms are intended to be inclusive, as is the term “comprising.” Furthermore, the term “or” as used in this disclosure is not intended to mean exclusive OR.

[0075] In this disclosure, if articles are added through translation, such as a, an, and the in English, this disclosure may include the fact that the noun following these articles is plural.

[0076] In this disclosure, the term "A and B are different" may mean "A and B are different from each other." The term may also mean "A and B are each different from C." Terms such as "separate" and "combine" may be interpreted similarly to "different." [Explanation of Symbols]

[0077] 1, 1A, 1B...Privacy protection data linkage system, 10, 10A, 10B, 10C...Devices, 10X, 10Y...Encryption devices, 10Z...Aggregation devices, 11...De-identification processing unit, 12...Encryption unit, 13...Aggregation processing unit, 14...Confidentiality processing unit, 1001...Processor, 1002...Memory, 1003...Storage, 1004...Communication device, 1005...Input device, 1006...Output device, 1007...Bus.

Claims

1. The system includes multiple devices that hold user data, including user IDs and personal information about the user. Each of the above devices is An anonymization processing unit that deidentifies the user data and generates anonymized data, An encryption unit performs a first encryption using its own encryption key on the anonymized data, and a second encryption using its own encryption key on the anonymized data received from the other party's device. An aggregation processing unit compares the second encrypted deidentified data with the second encrypted deidentified data received from the other party's device, and aggregates the number of second encrypted deidentified data where the user ID corresponding portion, identified based on predetermined structural information of the user data, matches, as the number of target users. Includes, The anonymization processing unit extracts the user data based on the conditions that the personal information must satisfy, and generates anonymized data corresponding to the extracted user data. The de-identification processing unit generates grouped de-identification data corresponding to groups of user data where the personal information satisfies each of the multiple conditions, if the conditions that the personal information must satisfy include multiple conditions. Privacy-protecting data sharing system.

2. The system includes multiple devices that hold user data, including user IDs and personal information about the user. Each of the above devices is An anonymization processing unit that deidentifies the user data and generates anonymized data, An encryption unit performs a first encryption using a first encryption key held by the first device, which is the device itself, on the anonymized data, and a second encryption using the first encryption key on the anonymized data that has been encrypted in the first encryption, which has been received from the second device. A third device comprises an aggregation processing unit that compares the second encrypted deidentified data from the first device with the second encrypted deidentified data from the second device, and aggregates the number of second encrypted deidentified data where the user ID corresponding portion, identified based on predetermined structural information of the user data, matches, as the number of target users. Includes, The anonymization processing unit extracts the user data based on the conditions that the personal information must satisfy, and generates anonymized data corresponding to the extracted user data. The de-identification processing unit generates grouped de-identification data corresponding to groups of user data where the personal information satisfies each of the multiple conditions, if the conditions that the personal information must satisfy include multiple conditions. Privacy-protecting data sharing system.

3. The system includes multiple encryption devices that store and encrypt user data, including user IDs and personal information about users, and an aggregation device. Each of the aforementioned encryption devices is An anonymization processing unit that deidentifies the user data and generates anonymized data, An encryption unit performs a first encryption using its own encryption key on the anonymized data, and a second encryption using its own encryption key on the anonymized data received from the other party's device. Includes, The aforementioned aggregation device is An aggregation processing unit compares the second encrypted deidentified data received from the encryption device with the second encrypted deidentified data received from the counterparty device of the encryption device, and aggregates the number of second encrypted deidentified data where the user ID corresponding portion, identified based on predetermined structural information of the user data, matches, as the number of target users. Includes, The anonymization processing unit extracts the user data based on the conditions that the personal information must satisfy, and generates anonymized data corresponding to the extracted user data. The de-identification processing unit generates grouped de-identification data corresponding to groups of user data where the personal information satisfies each of the multiple conditions, if the conditions that the personal information must satisfy include multiple conditions. Privacy-protecting data sharing system.

4. The anonymization processing unit further performs processing for privacy protection of the personal information. A privacy-protecting data linkage system according to any one of claims 1 to 3.

5. Each of the above devices is A confidentiality processing unit performs confidentiality processing on the aggregated results obtained by the aforementioned aggregation processing unit to generate confidential statistical information. The privacy-protecting data sharing system according to claim 1 or 2, further comprising:

6. The aforementioned aggregation device is A confidentiality processing unit performs confidentiality processing on the aggregated results obtained by the aforementioned aggregation processing unit to generate confidential statistical information. The privacy-protecting data linkage system according to claim 3, further comprising:

7. An anonymization processing unit that deidentifies user data including a user ID and personal information relating to the user and generates anonymized data, An encryption unit performs a first encryption using its own encryption key on the anonymized data, and a second encryption using its own encryption key on the anonymized data received from the other party's device. An aggregation processing unit compares the second encrypted deidentified data with the second encrypted deidentified data received from the other party's device, and aggregates the number of second encrypted deidentified data where the user ID corresponding portion, identified based on predetermined structural information of the user data, matches, as the number of target users. Includes, The anonymization processing unit extracts the user data based on the conditions that the personal information must satisfy, and generates anonymized data corresponding to the extracted user data. The de-identification processing unit generates grouped de-identification data corresponding to groups of user data where the personal information satisfies each of the multiple conditions, if the conditions that the personal information must satisfy include multiple conditions. Privacy-protecting data sharing device.

Images