Secure computation system, information processing system, secure computation method, and program
The secure computing system integrates multiple models using secure calculation methods to enhance predictive accuracy while maintaining model confidentiality, addressing the limitations of decentralized model integration.
Patent Information
- Application Number
- JP2023529413
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2021-06-25
- Publication Date
- 2024-11-13
- Estimated Expiration
- 2041-06-25
AI Technical Summary
Existing systems limit the accuracy of predictive models when using decentralized models, as individual service providers are reluctant to share their models due to confidentiality concerns, and integrating multiple models for improved accuracy is challenging.
A secure computing system that integrates multiple models generated by different service providers using secure calculation methods, such as homomorphic encryption and multi-party computation, to maintain model confidentiality while enhancing predictive accuracy.
The system provides more accurate predictive models by integrating multiple models while keeping each model's parameters secret, thus overcoming the limitations of individual model usage.
Smart Images

Figure 0007582473000001 
Figure 0007582473000002 
Figure 0007582473000003
Abstract
Description
[Technical field]
[0001] The present disclosure relates to a secure computation system, a business entity server, an information processing system, a secure computation method, and a recording medium. [Background technology]
[0002] Government agencies use personal information of residents and others that they hold to analyze ways to solve problems in the local governments they supervise. Models using AI (Artificial Intelligence) are used as a means of analyzing ways to solve problems. To improve the performance and accuracy of the models, technology is used that performs predictive processing on distributed personal information data while protecting privacy.
[0003] For example, Patent Document 1 discloses a system that performs prediction processing while still encrypted, using an encrypted prediction model and user information encrypted in a similar manner to the distributed prediction model. [Prior art documents] [Patent documents]
[0004] [Patent Document 1] JP 2019-215512 A Summary of the Invention [Problem to be solved by the invention]
[0005] However, the invention described in the above-mentioned Patent Document 1 outputs data predicted by a single prediction model, so there is a limit to how much accuracy can be improved in the prediction data. When a service provider entrusted with a business by a city's local government analyzes personal information of residents, a more accurate model can be generated by using information on models held by multiple service providers, rather than using models held by individual service providers. However, since the models held by each service provider are information that should be treated as know-how, there are cases where the service provider wishes to keep the information secret.
[0006] One example of the objective of the present disclosure is to provide more accurate models while keeping each model confidential. [Means for solving the problem]
[0007] A secure computing system in one aspect of the present disclosure is used for analyzing optimization methods for a city based on personal data of individuals belonging to the city, and includes a parameter receiving means for receiving input of a plurality of anonymized parameters of a plurality of models generated by each of a plurality of businesses, a secure computing means for integrating the aonymized parameters through secure computation, and an output means for outputting the parameters integrated by the secure computing means in a concealed format.
[0008] The business server in one aspect of the present disclosure comprises a model storage means for storing a model that analyzes optimization methods for a city based on personal data of individuals belonging to the city, an anonymization means for anonymizing parameters of the model stored in the model storage means, a model input / output means for transmitting the parameters of the model to a secure computing system in an anonymized form, a restoration means for restoring the anonymized parameters, and an analysis means for performing analysis on the optimization of the city using an updated model updated by federated learning using secure computing based on information held by each city, and the analysis means outputs proposed actions for solving the city's problems based on the personal data of residents belonging to the city.
[0009] An information processing system in one aspect of the present disclosure is an information processing system having a plurality of business operators servers and a secure computing system, wherein the plurality of business operators servers are each generated based on personal data of individuals belonging to each city, and include a model storage means for storing a model to be analyzed for an optimization method for the city, an anonymization means for anonymizing parameters of the model stored in the model storage means, a model input / output means for transmitting the model to the secure computing system in an anonymized format, and a restoration means for restoring the anonymized parameters, and the secure computing system is used to analyze an optimization method for the city based on the personal data of individuals belonging to the city, and includes a parameter receiving means for receiving input of a plurality of anonymized parameters of a plurality of models generated by each of the plurality of business operators, a secure computing means for integrating the aonymized parameters by secure computation, and an output means for outputting the parameters integrated by the secure computing means in an anonymized format.
[0010] A secure computation method in one aspect of the present disclosure is used to analyze optimization methods for a city based on personal data of individuals belonging to the city, and receives input of a plurality of anonymized parameters of a plurality of models generated by each of a plurality of businesses, integrates the anonymized parameters using secure computation, and outputs the integrated parameters in anonymized form.
[0011] A recording medium in one aspect of the present disclosure stores a program that causes a computer to execute the following steps: (1) analyzing an optimization method for a city based on personal data of individuals belonging to the city, (2) receiving input of a plurality of anonymized parameters of a plurality of models generated by each of a plurality of businesses, (3) integrating the anonymized parameters through secret computation, and (4) outputting the integrated parameters in anonymized format. Effect of the Invention
[0012] One example of an effect of the present disclosure is the ability to provide more accurate models while keeping each model confidential. [Brief description of the drawings]
[0013] [Figure 1] FIG. 1 is a block diagram showing the configuration of an information processing system according to the first embodiment. [Diagram 2] FIG. 2 is a diagram for explaining a service provider that implements the information processing system in the first embodiment. [Diagram 3] FIG. 3 is a diagram showing a hardware configuration in which the secure computation system according to the first embodiment is realized by a computer device and its peripheral devices. [Figure 4] FIG. 4 is a flowchart showing the operation of secure computation in the first embodiment. [Diagram 5] FIG. 5 is a block diagram showing a configuration of an information processing system according to the second embodiment. [Figure 6] FIG. 6 is a flowchart showing an information processing operation in the second embodiment. [Figure 7] FIG. 7 is a block diagram illustrating a configuration of a trustee selection unit in a modified example of the second embodiment. [Figure 8] FIG. 8 is a flowchart showing the operation of selecting a trustee in the modified example of the second embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0014] Next, an embodiment will be described in detail with reference to the drawings.
[0015] [First embodiment] FIG. 1 is a block diagram showing a configuration of an information processing system 10 according to the first embodiment.
[0016] Referring to FIG. 1, the information processing system 10 includes a secure computation system 100 and a plurality of business servers 200 (200a, 200b). The secure computation system 100 includes a parameter receiving unit 101, a secure computation unit 102, and an output unit 103. The secure computation system 100 is implemented by a model federation business that provides each service business with an analysis tool for a city optimization method, etc. Each of the business servers 200 includes a model storage unit 201 (201a, 201b) that stores a trained model for analyzing personal information (personal data) of residents or information related to city optimization, an anonymization unit 202 (202a, 202b) that anonymizes parameters, a model input / output unit 203 (203a, 203b) that inputs and outputs parameters to and from the secure computation system 100, and a restoration unit 204 (204a, 204b) that restores the anonymized parameters. In this embodiment, the number of the business servers 200 is two, but is not limited to this. The number of the business servers 200 is the same as the number of the service businesses that perform federated learning. The secure computation system 100, which is an essential component of this embodiment, will be described in detail below.
[0017] FIG. 2 is a diagram for explaining a service provider implementing the information processing system 10 in the first embodiment. As shown in FIG. 2, as an example of the first embodiment, each of service providers A and B is entrusted with a project related to the issues of the cities by the local government that governs cities A and B, and cities C and D, and is granted the right to use the personal information of the residents belonging to the entrusted cities. In this case, in this embodiment, each of service providers A and B generates a model for analyzing the issues of the cities using the information of the residents received from each city. Next, service providers A and B transmit the parameters of the model to the model federation provider in a confidential format. Then, the model federation provider that receives the multiple confidential parameters integrates the multiple parameters. As one aspect of this embodiment, the model federation provider integrates the models generated by each service provider and is used to generate a more accurate model ((1) in FIG. 2). In FIG. 2, the direction of the arrow between the service provider and the model federation provider indicates the direction in which information regarding the parameters is transmitted. The model federation operator obtains the anonymized parameters of the model from service operators A and B, and transmits the integrated parameters to each service operator in anonymized format.
[0018] As an example of another aspect of the first embodiment, parameters of models held by service providers A and B are integrated by a model federation operator using secret computation to generate a more accurate model ((2) in Figure 2).
[0019] Fig. 3 is a diagram showing an example of a hardware configuration in which the secure computation system 100 in the first embodiment of the present disclosure is realized by a computer device 500 including a processor. As shown in Fig. 3, the secure computation system 100 includes a CPU (Central Processing Unit) 501, memories such as a ROM (Read Only Memory) 502 and a RAM (Random Access Memory) 503, a storage device 505 such as a hard disk for storing a program 504, a communication I / F (Interface) 508 for network connection, and an input / output interface 511 for inputting and outputting data. In the first embodiment, parameter information received from each business entity server 200 is input to the secure computation system 100 via the input / output interface 511.
[0020] The CPU 501 operates an operating system to control the entire secure computation system 100 according to the first embodiment of the present invention. The CPU 501 also reads out programs and data from a recording medium 506 mounted in, for example, a drive device 507 into a memory. The CPU 501 also functions as the parameter receiving unit 101, the secure computation unit 102, and the output unit 103 in the first embodiment, or as a part of these, and executes processes or commands in the flowchart shown in Fig. 4, which will be described later, based on the program.
[0021] The recording medium 506 is, for example, an optical disk, a flexible disk, a magneto-optical disk, an external hard disk, or a semiconductor memory. A recording medium that is a part of the storage device is a non-volatile storage device, and the program is recorded therein. The program may also be downloaded from an external computer (not shown) that is connected to a communication network.
[0022] The input device 509 is realized by, for example, a mouse, a keyboard, built-in key buttons, etc., and is used for input operations. The input device 509 is not limited to a mouse, a keyboard, or built-in key buttons, and may be, for example, a touch panel. The output device 510 is realized by, for example, a display, and is used to check the output.
[0023] As described above, the first embodiment shown in FIG. 1 is realized by the computer hardware shown in FIG. 3. However, the means for realizing each part of the secure computation system 100 in FIG. 1 is not limited to the configuration described above. The secure computation system 100 may be realized by one physically combined device, or may be realized by two or more physically separated devices connected by wire or wirelessly. For example, the input device 509 and the output device 510 may be connected to the computer device 500 via a network. The secure computation system 100 in the first embodiment shown in FIG. 1 may also be configured by cloud computing or the like.
[0024] In FIG. 1, the parameter receiving unit 101 is a means for receiving input of a plurality of concealed parameters of a plurality of models used in the analysis of the optimization method of a city. The plurality of models are generated in each of a plurality of businesses based on information of individuals belonging to each of the cities. The parameter receiving unit 101 receives and receives the parameters of the trained model in a concealed format in each of the plurality of business servers 200 via the communication I / F 508 through the network, triggered, for example, by an operation for integrating parameters by a model association business operator. The trained model is a model that is determined in advance by machine learning in each business server 200 using, for example, personal data of residents or commissioned business information from the government in order to output a specific analysis result. The machine learning model includes, but is not limited to, a decision tree model, a linear regression model, a logistic regression model, a neural network model, and the like.
[0025] Personal data is personal data of a specific individual, and includes, for example, attribute information, health information, happiness level (well-being level), behavior history information, or personal information collected from a sensor of an individual. Personal data is obtained by converting questionnaires, health checkup results, or sensing data obtained from an individual into attribute information, health information, well-being level, behavior history, situation, or condition, etc. Attribute information is, for example, age or sex. Health information is, for example, information indicating physical build such as height and weight, or information on the results of a health checkup. Well-being level is, for example, information from which an individual's lifestyle needs and happiness level can be extracted based on questionnaire results, etc. The situation is, for example, information from which an individual's recent situation can be grasped, which is obtained from the contents of a report to an administrative agency. The condition is information obtained from sensing data or behavior history data. These pieces of information are, for example, acquired through a network from a mobile terminal possessed by an individual or sensors and cameras installed in the city.
[0026] The secure computation unit 102 is a means for integrating, by secure computation, a plurality of parameters that have been received by the parameter receiving unit 101 and made anonymous. In this embodiment, integrating a plurality of parameters that have been made anonymous by secure computation means that the secure computation system 100 performs machine learning in a state where the secure computation system 100 is distributed to each business entity server 200 (federated learning), and integrates the parameters of a model that has already been learned in each business entity server 200 by using secure computation. In this embodiment, this also includes the secure computation system 100 integrating the parameters of a model that each business entity server 200 has machine-learned.
[0027] The secure computing unit 102 integrates the concealed parameters according to a predetermined combination rule. A known method can be used as a parameter integration method, and for example, when integrating, the weight of the parameters corresponding to each model can be changed according to the characteristics of each model.
[0028] As the secure computation method, a special encryption corresponding to a specific process such as homomorphic encryption, a trusted execution environment (Trusted Execution Environment) in which processing is performed in an isolated state on hardware, or multi-party computation in which computation is performed (secret sharing computation) while keeping secret sharing among multiple servers can be used. Specific examples of the secure computation method of multi-party computation include the following. For example, the concealed data a is secretly shared into shared values x, y, ..., and x, y, ... are sent to different servers by administrators. Next, while keeping the concealed data a secretly shared, the computation proceeds while communicating with each other, and finally, the shared values u, v, ... of the output, which are the computation results of each server, are collected and restored, thereby obtaining the computation result F(a). This computation result becomes a parameter that integrates the parameters of each model. For this reason, when multi-party computation is used as the secure computation method, the secure computation unit 102 includes multiple servers. According to multi-party computation, management of encryption keys and an isolated environment are not required, and computation processing is faster. The secure computation unit 102 outputs the model parameters thus obtained to the output unit 103 in a concealed format.
[0029] The output unit 103 is a means for transmitting the parameters integrated by the secure computation unit 102 to the business server 200. The output unit 103 transmits the integrated parameters in a format that allows the business server 200 to update the parameters of the model. When transmitting to the business server 200, the output unit 103 can transmit the differences of the updated parameters (only the improvements) instead of the updated parameters.
[0030] The operation of the secure computation system 100 configured as above will now be described with reference to the flow chart of FIG.
[0031] 4 is a flowchart showing an outline of the operation of the secure computing system 100 in the second embodiment. Note that the processing according to this flowchart may be executed based on program control by the above-mentioned processor.
[0032] As shown in Fig. 4, first, the parameter receiving unit 101 receives input of a plurality of concealed parameters for each of models that have been trained in a plurality of business operators' servers 200 (step S101). Next, the secure computation unit 102 integrates the plurality of concealed parameters by secure computation (step S102). Finally, the output unit 103 outputs the model parameters calculated by the secure computation unit 102 in a concealed format (step S103). With this, the secure computation system 100 ends the operation of the secure computation.
[0033] In the secure computation system 100, the secure computation unit 102 integrates a plurality of concealed parameters through secure computation, thereby making it possible to provide a more accurate model while keeping the parameters of each model secret.
[0034] [Second embodiment] Next, a second embodiment of the present disclosure will be described in detail with reference to the drawings. Hereinafter, the contents overlapping with the above description will be omitted to the extent that the description of this embodiment is not unclear. In the second embodiment, the information processing system 11 is used to provide an updated model by federated learning using secure computation. These updated models are used, for example, as a tool for analyzing an optimization method for a city toward realizing a smart city. As with the computer device shown in FIG. 3, the functions of each component in each embodiment of the present disclosure can be realized not only by hardware but also by a computer device or firmware based on program control. These models are models generated based on, for example, individual problems (requests) or city problem information and actions and plans for solving those problems, and are stored in each model storage unit 211 (211a, 211b) of each business operator server 210 (210a, 210b).
[0035] FIG. 5 is a block diagram showing a configuration of an information processing system 11 including a secure computation system 110 according to a second embodiment of the present disclosure. With reference to FIG. 5, the secure computation system 110 and the business entity server 210 (210a, 210b) according to the second embodiment will be described, focusing on the parts different from the information processing system 10 according to the first embodiment. The secure computation system 110 according to the second embodiment includes a parameter receiving unit 111, a secure computation unit 112, and an output unit 113. The multiple business entity servers 210 (210a, 210b) include a model storage unit 211 (211a, 211b), a concealment unit 212 (212a, 212b), a model input / output unit 213 (213a, 213b), a restoration unit 214 (214a, 214b), and an analysis unit 215 (215a, 215b). The secure computation system 110 integrates multiple parameters of the trained models received from multiple business servers 210a, 210b using secure computation.
[0036] Here, we will explain the details of each model used to analyze the optimization method for a city based on information of individuals belonging to that city. In this embodiment, optimization of a city means, for example, solving urban problems. The optimization method for a city is a method for solving urban problems.
[0037] In this embodiment, a city refers to, for example, an area governed by a specific administrative agency and the administrative agency, and includes not only areas with a high population density but also rural areas. Urban issues are indicated as performance indicators or optimization goals, which are index values for quantitatively grasping the results (degree of achievement) of projects corresponding to administrative issues set by each city. Administrative issues include, for example, improving the health of residents, economic development of the city, and environmental issues. For example, if the administrative issue is improving the health of residents, performance indicators such as a 10% reduction in nursing care insurance premiums or a 20% reduction in medical expenses are set.
[0038] The optimization goal is a specific measure to achieve the performance indicator, and is information that, when input into the trained model, can analyze the proposed actions to execute the measure. The proposed actions are actions recommended to each individual to solve the urban issues. The proposed actions are, for example, actions that solve the urban issues and also satisfy the individual requirements. The individual requirements are, for example, requirements analyzed based on personal data obtained from the individual.
[0039] For example, if the optimization goal is to improve the health of residents, the proposed action may be to have the residents walk a specific distance or to improve the numerical value of a specific item in the health check results. By integrating multiple parameters of models that have already been trained by each service provider, these models can be updated to parameters that reflect the analysis results for each city, thereby improving the accuracy of the models. Here, the models used in this embodiment will be described.
[0040] An example of a model used in this embodiment is a model generated by learning personal data of an individual and a classification name classified based on the personal data as learning data. Examples of methods for classifying individuals include classification by attributes (age, sex), classification by foods that can be ingested based on health check results, and classification by the amount of exercise required based on exercise history such as walking distance.
[0041] Another example of a model used in this embodiment is a model that specifies and outputs a suggested action for an individual by inputting classification information into a trained model. For example, in the learning process, this model generates a trained model for each combination that indicates the relationship between one or more combinations of classification and optimization goal acquired as learning data and an action (that achieves the request and optimization goal) that indicates the correct label of the learning data, using a neural network, graph AI, or other machine learning algorithm. During learning, the trained model may be updated and strengthened by verifying the trained model based on the proposal acceptance rate from the individual when the suggested action is actually presented to the individual.
[0042] In addition, in the estimation process, when classification information is input, the contents of the proposed action that satisfies the request and the optimization goal are estimated using a trained model corresponding to the combination of the classification and the optimization goal. As described above, the model is trained using the training data, and the contents of the proposed action are specified. For example, when an individual's request is a healthy lifestyle, the model is a model that outputs a recommended list of menus when ingestible foods (e.g., daily salt amount, etc.) are input as classification information. Another example of the model is a model that outputs a recommended list of restaurants when the required amount of exercise, individual location information, or restaurant location information is input to the model. Still another example of the model is a model that outputs a recommended menu from a restaurant when a menu at the restaurant is input. In this model, menus at one restaurant or multiple restaurants may be input.
[0043] <Secure computation system> The secure computation unit 112 receives the parameters of the trained model of each service provider from the business server 210 through the input / output interface 511. Next, the secure computation unit 112 integrates the received multiple concealed parameters by secure computation according to a predetermined combination rule, and outputs the parameters of the integrated model to the output unit 113 in a concealed format. The output unit 113 transmits the parameters of the integrated model to each business server 210 through the model input / output unit 213. In addition, after transmitting the parameters to the business server 210, if the business server 210 again learns the model and updates the parameters, the secure computation system 110 may receive the updated parameters again. Note that the operations of the parameter reception unit 111, the secure computation unit 112, and the output unit 113 are similar to the operations of the parameter reception unit 101, the secure computation unit 102, and the output unit 103 in the first embodiment, and therefore will not be described here.
[0044] <Business server> The business entity server 210 updates the model stored in the model storage unit 211 to a model to which the parameters received from the secure computation system 110 are applied. Specifically, the model input / output unit 213 receives the parameters in a format that is still concealed, and outputs them to the restoration unit 214. Next, the restoration unit 214 restores the parameters and replaces them with the parameters of the model stored in the model storage unit 211. Next, the analysis unit 215 performs analysis using the updated model. The updated model is incorporated in a tool for analyzing a city optimization method used by a service provider. The analysis unit 215 performs analysis of city issues using the updated model, triggered by a user's operation on the analysis tool, and outputs the analysis results in a state that can be viewed on a display device or the like. For example, in the tool for analyzing a city optimization method, when information on personal data of multiple individuals belonging to each city is input, the analysis unit 215 classifies the individuals based on the personal data and outputs the classification name. Furthermore, when the analysis unit 215 receives classification information such as a classification name classified based on the personal data of a plurality of individuals belonging to a city, it outputs a suggested action to be proposed to individuals of that classification. In order to improve the accuracy of the analysis results by the analysis unit 215, the business entity server 210 may perform learning again based on additionally obtained personal data and transmit the updated parameters to the secure computation system 110. In this way, the accuracy of the model can be further improved by repeating the updating of parameters through learning in each business entity server 210 and the integration of parameters in the secure computation system 110 until, for example, a predetermined condition is satisfied.
[0045] The operation of the information processing system 11 configured as above will be described with reference to the flowchart of FIG.
[0046] 6 is a flowchart showing an outline of the operation of the information processing system 11 in the first embodiment. Note that the processing according to this flowchart may be executed based on program control by the above-mentioned processor.
[0047] As shown in FIG. 6, first, the business entity server 210 locally learns using personal information of residents held by the service entity (step S201). Next, the anonymization unit 212 anonymizes the parameters of the model learned in each business entity server 210 (step S202). Next, the model input / output unit 213 outputs the parameters in anonymized format to the secure computation system 110 (step S203). Next, the parameter reception unit 111 of the secure computation system 110 receives the anonymized parameters (step S204). Next, the secure computation unit 112 integrates the multiple anonymized parameters by secure computation (step S205). Next, the output unit 113 outputs the parameters integrated by the secure computation unit 112 in anonymized format to each of the business entity servers 210 (step S206). Next, the business entity server 210 acquires the integrated parameters in anonymized format through the model input / output unit 213 (step S207). Next, the restoration unit 214 restores the concealed parameters (step S208). Next, the business server 210 updates the model stored in the model storage unit 211 to a model to which the restored parameters are applied (step S209). Next, the business server 210 judges whether or not a predefined condition is satisfied (step S210). If the predefined condition is satisfied (step S210; YES), the analysis unit 215 starts analysis using the updated model and ends the flow (step S211). If the predefined condition is not satisfied, the business server 210 returns to step S201 (step S210; NO) and executes the flow again. With the above, the information processing system 11 ends the operation of the secure computation.
[0048] In the second embodiment of the present disclosure, parameters of multiple models for analyzing urban optimization methods are integrated to output more accurate analysis results.
[0049] [Modification of the second embodiment] A modified example of the second embodiment will be described. In the modified example of the second embodiment, the multiple businesses of the second embodiment are candidates for contractors to which the city's local government will entrust the implementation of the optimization method. In addition, the modified example of the second embodiment further includes a contractor selection unit 119 for selecting contractor candidates. In the modified example of the second embodiment, for example, it is assumed that the activities of the local government are entrusted to a private company by a performance-linked private contracting method (PFS: Pay For Success) or the like. That is, it is assumed that the private company carries out activities to achieve the performance indicators that are the city's goals set by the local government. The contractor selection unit 119 matches the business entrusted by the government with the company that will receive the business.
[0050] Fig. 7 is a block diagram showing a configuration of the trustee selection unit 119 in a modified example of the second embodiment. As shown in Fig. 7, the trustee selection unit 119 includes a business information receiving unit 1191 that receives input of information related to the entrusted business, a trustee candidate extraction unit 1192 that extracts trustee candidates from past performance information of the business related to the entrusted business, and a trustee specification unit 1193 that specifies a trustee from the trustee candidates extracted by the trustee candidate extraction unit 1192.
[0051] The business information receiving unit 1191 receives input of information related to the commissioned business through the input device 509. The information related to the commissioned business is, for example, the commissioned business period, the performance indicators, and the success fee amount according to the achievement level of the performance indicators. The success fee amount may be set in stages according to the achievement level of the performance indicators. The success fee amount may be higher as the achievement level becomes higher, for example, 10 million yen for a 10% reduction in medical expenses, and 15 million yen for a 15% reduction in medical expenses.
[0052] The candidate contractor extraction unit 1192 extracts, via a network, information on company data (candidate contractors) having past performance related to the performance indicators received by the business information reception unit 1191. The candidate contractor extraction unit 1192 may extract, for example, past performance information from administrative document management information registered in a blockchain between multiple administrative agencies.
[0053] The trustee specification unit 1193 specifies a trustee based on the past performance of the trustee candidates extracted by the trustee candidate extraction unit 1192 and the evaluation information on that performance. The evaluation information includes, for example, the achievement level of the performance indicators and the presence or absence of problems at the time of past entrustment. The trustee specification unit 1193 specifies a trustee from among the trustee candidates using a trustee analysis model generated based on the contents of the past performance and the evaluation information on that performance. This model is, for example, a model that specifies and outputs the optimal trustee from among the trustee candidates when information on the trustee candidates extracted by the trustee candidate extraction unit 1192 is input. This model is, for example, a model generated by a decision tree, a neural network, a regression model, or a deep learning neural network, and is stored in the storage device 505. In addition, in this embodiment, a model that outputs the optimal trustee when information on the entrusted business is input may be used. In this case, a series of operations is automatically executed, including the reception of input information on the entrusted business by the business information reception unit 1191, the extraction of the trustee candidates by the trustee candidate extraction unit 1192, and the specification of the trustee by the trustee specification unit 1193. The trustee specification unit 1193 outputs information on the trustee thus specified by the output device 510, for example.
[0054] In addition, in a modified example of this embodiment, after the commissioned work is completed, the automatic calculation and payment of the success fee may be performed by a smart contract. A smart contract is a mechanism that is executed on a blockchain network and is executed to perform a specific operation when a specific condition is satisfied as a trigger. The payment of the success fee using a smart contract is performed by the commissioned business entity inputting the results of the performance indicators of the commissioned business into the blockchain, and the amount of the success fee is automatically calculated, and the calculated amount of the success fee is paid to the commissioned business entity.
[0055] The operation of the outsourcing selection unit 119 configured as above will be described with reference to the flowchart in Fig. 8. First, the business information receiving unit 1191 receives input of information related to the outsourcing business (step S211), then the outsourcing candidate extraction unit 1192 extracts information of company data having past performance related to the received performance indicator as outsourcing candidates (step S212). Finally, the outsourcing specification unit 1193 inputs the extracted information of the outsourcing candidates into a model and specifies the outsourcing company (step S213). With this, the outsourcing selection unit 119 ends the outsourcing selection operation.
[0056] In a modified example of this embodiment, a contractor is identified using a contractor analysis model created based on past performance details and evaluation information on those performances, thereby making it possible to select an optimal contractor.
[0057] Although the present invention has been described above with reference to the embodiments, the present invention is not limited to the above-mentioned embodiments. Various modifications that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention.
[0058] For example, although a number of operations are described in a sequence in the form of a flowchart, the sequence does not limit the order in which the operations are performed. Therefore, when implementing each embodiment, the sequence of the operations can be changed as long as it does not affect the content.
[0059] A part or all of the above-described embodiments can be described as, but is not limited to, the following supplementary notes.
[0060] (Appendix 1) A parameter receiving means for receiving an input of a plurality of anonymized parameters of a plurality of models generated by a plurality of businesses, the plurality of models being used in an analysis of an optimization method for the city based on personal data of individuals belonging to the city; a secure computation means for integrating the received multiple anonymized parameters through secure computation; an output means for outputting the parameters integrated by the secure computation means in a concealed form; A secure computing system.
[0061] (Appendix 2) The secure computing system of claim 1, wherein the model is a model that classifies individuals based on their personal data when information regarding the personal data of the individuals belonging to each of the cities is input.
[0062] (Appendix 3) The secure computing system of claim 1 or 2, wherein the model is a model that, when classification information classified based on the personal data of the individuals belonging to the city is input, outputs suggested actions to be proposed to individuals of that classification.
[0063] (Appendix 4) 4. The secure computation system according to claim 1, wherein the secure computation is a secret sharing computation.
[0064] (Appendix 5) The secure computation system according to any one of appendices 1 to 4, wherein the plurality of businesses in the secure computation means are candidates for contractors to which the implementation of the optimization method will be entrusted by a local government of the city.
[0065] (Appendix 6) The secure computing system according to claim 5, further comprising a trustee selection means for selecting the candidate trustees based on past performance information of businesses related to the business to be entrusted.
[0066] (Appendix 7) The secret computing system described in Appendix 6, wherein the trustee selection means includes a business information receiving means for receiving input of information regarding the entrusted business, a candidate trustee extraction means for extracting candidate trustees from past performance information of businesses related to the business to be entrusted, and a candidate trustee identification means for identifying a trustee from among the candidate trustees extracted by the candidate trustee extraction means.
[0067] (Appendix 8) The secure computing system according to claim 7, wherein the candidate trustee extraction means acquires past performance information based on administrative document management information.
[0068] (Appendix 9) The secure computing system according to claim 7 or 8, wherein the trustee identification means identifies the trustee using a model generated based on past performance and evaluation information on the performance.
[0069] (Appendix 10) A model storage means for storing a model for analyzing an optimization method for a city based on personal data of individuals belonging to the city; an anonymization means for anonymizing parameters of the model stored in the model storage means; a model input / output means for transmitting the parameters of the model to a secure computing system in a concealed format; A restoration means for restoring the concealed parameters; and an analysis means for performing an analysis on the optimization of each city by using an updated model updated by federated learning using secure computation based on information held by each city, The analysis means is a business server that outputs proposed actions for solving the problems of the city based on personal data of residents belonging to the city.
[0070] (Appendix 11) An information processing system having a plurality of business servers and a secure computing system, Each of the plurality of business entity servers has a model storage means for storing a model for analyzing an optimization method for the city, the model being generated based on personal data of individuals belonging to the city; an anonymization means for anonymizing parameters of the model stored in the model storage means; a model input / output means for transmitting the model in a concealed format to a secure computing system; A restoration means for restoring the concealed parameters; Equipped with The secure computing system comprises: A parameter receiving means for receiving an input of a plurality of anonymized parameters of a plurality of models generated by a plurality of businesses, the plurality of models being used in an analysis of an optimization method for the city based on personal data of individuals belonging to the city; a secure computation means for integrating the multiple parameters that have been made anonymous through secure computation; an output means for outputting the parameters integrated by the secure computation means in a concealed form; An information processing system comprising:
[0071] (Appendix 12) receiving an input of a plurality of anonymized parameters of a plurality of models generated by a plurality of businesses, the plurality of models being used for analyzing an optimization method for the city based on personal data of individuals belonging to the city; The multiple concealed parameters are integrated by secure computation; and outputting the integrated parameter.
[0072] (Appendix 13) receiving an input of a plurality of anonymized parameters of a plurality of models generated by a plurality of businesses, the plurality of models being used for analyzing an optimization method for the city based on personal data of individuals belonging to the city; The multiple concealed parameters are integrated by secure computation; A recording medium storing a program for causing a computer to execute the process of outputting the integrated parameters. [Explanation of symbols]
[0073] 10, 11 Information Processing Systems 100, 110 Secure Calculation System 101, 111 Parameter reception section 102, 112 Secret Computation Department 103, 113 Output section 200, 210 Business server 201, 211 Model memory section 202, 212 Concealment Department 203, 213 model input / output section 204, 214 Restoration Department 215 Analysis Department
Claims
1. A parameter receiving means for receiving an input of anonymized parameters of a plurality of models used in an analysis of an optimization method for a city based on personal data of individuals belonging to the city and generated by each of a plurality of businesses; a secure computation means for integrating the multiple concealed parameters through secure computation; an output means for outputting the parameters integrated by the secure computation means in a concealed format; A means for selecting a trustee; Equipped with the plurality of businesses in the secure computing means are candidates for contractors to be entrusted with the implementation of the optimization method by a local government of the city, The outsourcing selection means selects the outsourcing candidates based on past performance information of businesses related to the business to be outsourced. Secret calculation system.
2. 2. The secure computing system according to claim 1, wherein the model is a model that, when information regarding the personal data of the individuals belonging to each of the cities is input, classifies the individuals based on the personal data.
3. The secure computing system of claim 1 or 2, wherein the model is a model that, when classification information classified based on the personal data of the individuals belonging to the city is input, outputs suggested actions to be proposed to the individuals of the classification.
4. 4. The secure computation system according to claim 1, wherein the secure computation is a secret sharing computation.
5. The trustee selection means includes a business information receiving means for receiving input of information regarding the entrusted business, a candidate trustee extraction means for extracting candidate trustees from past performance information of businesses related to the business to be entrusted, and a candidate trustee identification means for identifying a trustee from among the candidate trustees extracted by the candidate trustee extraction means.A secure computing system as described in any one of claims 1 to 4.
6. The entrustment candidate extraction means acquires past performance information based on administrative document management information. The secure computing system according to claim 5.
7. 7. The secure computing system according to claim 5, wherein the trustee specification means specifies the trustee by using a model generated based on past performance and evaluation information on the performance.
8. An information processing system having a plurality of business servers and a secure computing system, Each of the plurality of business entity servers has a model storage means for storing a model for analyzing an optimization method for the city, the model being generated based on personal data of individuals belonging to the city; an anonymization means for anonymizing parameters of the model stored in the model storage means; a model input / output means for transmitting the parameters of the model to a secure computing system in a concealed format; A restoration means for restoring the concealed parameters; Equipped with The secure computing system comprises: A parameter receiving means for receiving an input of a plurality of anonymized parameters of a plurality of models generated by a plurality of businesses, the plurality of models being used in an analysis of an optimization method for the city based on personal data of individuals belonging to the city; a secure computation means for integrating the multiple parameters that have been made anonymous through secure computation; an output means for outputting the parameters integrated by the secure computation means in a concealed format; A means for selecting a trustee; Equipped with the plurality of businesses in the secure computing means are candidates for contractors to be entrusted with the implementation of the optimization method by a local government of the city, The outsourcing selection means selects the outsourcing candidates based on past performance information of businesses related to the business to be outsourced. Information processing system.
9. receiving an input of a plurality of anonymized parameters of a plurality of models generated by a plurality of businesses, the plurality of models being used for analyzing an optimization method for the city based on the personal data of individuals belonging to the city; The concealed multiple parameters are integrated by secure computation; outputting the integrated parameters in a concealed form; The plurality of businesses are candidates for contractors to be commissioned by the city's local government to implement the optimization method, Selecting the candidate contractors based on past performance information of the business to be entrusted and the business related thereto; Secret calculation method.
10. receiving an input of a plurality of anonymized parameters of a plurality of models generated by a plurality of businesses, the plurality of models being used for analyzing an optimization method for the city based on the personal data of individuals belonging to the city; The concealed parameters are integrated by secure computation; outputting the integrated parameters in a concealed format; The plurality of businesses are candidates for contractors to be commissioned by the city's local government to implement the optimization method, Further, the computer is caused to select the candidate contractors based on past performance information of the businesses related to the business to be entrusted. program.
Citation Information
Patent Citations
Data sharing method based on collaborative deep learning
CN109977694A
Software procurement supporting system, software procurement supporting device, software procurement supporting program and computer-readable recording medium for recording software procurement supporting program
JP2005339155A
Vendor selection support system
JP2008077550A
Prediction model distribution method and prediction model distribution system
JP2019215512A
Health assistance system, information providing sheet output device, method, and program
WO2019187933A1
