JP7738432B2 - Information processing method, information processing device, and computer program - Google Patents

Description

The present invention relates to an information processing method, an information processing device, and a computer program.

Digital twins, which link real space (the real world) with virtual space (cyberspace), are attracting attention. Digital twins are a new technology that utilizes IoT (Internet of Things), which connects all things to the internet and allows them to communicate; artificial intelligence (AI); AR (Augmented Reality), which augments the real world by providing information; and VR (Virtual Reality), which can make virtual space appear more like the real world.

Patent Document 1 discloses a technology for assessing risks associated with vehicle operation by constructing a digital twin of an actual vehicle and updating the state of the digital twin in real time based on sensor data recorded by onboard sensors in the actual vehicle.

Japanese Patent Application Laid-Open No. 2020-13557

Digital twins, such as those in Patent Document 1, are increasingly being used primarily in the manufacturing industry, including automobiles and factory production equipment, because they enable real-time data collection and analysis using actual sensors, risk assessment, and identification of the causes of failures and errors. However, when it comes to digital twins of people, because each person's range of activities is different and their social lives span a wide range of activities, it is difficult to recreate a person's inner thoughts, values, and so on.

The present invention was made in consideration of these circumstances, and aims to provide an information processing method, information processing device, and computer program that can construct a digital twin of a person.

This application includes multiple means for solving the above problems, but one example is an information processing method that collects survey responses from the monitors themselves and uses the collected responses as learning data to adjust the monitor's digital twin.

This invention makes it possible to construct a digital twin of a person.

FIG. 1 illustrates an example of a configuration of an information processing system. FIG. 10 is a diagram showing an example of a questionnaire response from a monitor member. FIG. 1 illustrates an example of a configuration of an information processing device. FIG. 10 is a diagram illustrating an example of constructing a digital twin for each monitor member. FIG. 10 is a diagram showing an example of tuning the monitor's own digital twin. FIG. 1 is a diagram showing evaluation indicators included in Spranger's value typology (first evaluation axis). FIG. 10 is a diagram showing an example of a questionnaire regarding a first evaluation axis. FIG. 1 is a diagram showing the evaluation indicators included in the International Classification of Functioning, Disability and Health (second evaluation axis). FIG. 10 is a diagram showing an example of a questionnaire regarding a second evaluation axis. FIG. 1 is a diagram showing the degree of conformance in the evaluation axis of a digital twin. FIG. 10 is a diagram showing the degree of conformance in the evaluation index of a digital twin. FIG. 10 is a diagram showing an example of a classification table for a questionnaire. FIG. 10 is a diagram illustrating a processing procedure for tuning a digital twin by an information processing device. FIG. 2 is a diagram illustrating an example of the configuration of a questionnaire server. FIG. 10 is a diagram illustrating an example of a survey method using a survey server. FIG. 1 is a diagram illustrating an example of the configuration of a digital twin DB. FIG. 10 is a diagram illustrating an example of a questionnaire design screen. FIG. 2 is a diagram illustrating an example of the configuration of a questionnaire DB. FIG. 10 is a diagram showing an example of a monitor member response screen. FIG. 10 is a diagram showing an example of a processing procedure of a questionnaire method by a questionnaire server.

(First embodiment)
An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a diagram illustrating an example of the configuration of an information processing system. The information processing system includes a survey server 30 and an information processing device 50. The information processing system may also include a data server 100. The survey server 30, the information processing device 50, and the data server 100 are connected to a communication network 1. The data server 100 includes a survey DB 110 and is connected to a monitor member DB 120. The monitor member DB 120 records the attributes, email addresses, etc. of monitor members for each monitor member ID. The survey server 30 is connected to a terminal device 41 used by a person in charge of the research company. The number of terminal devices 41 is not limited to one. The survey server 30, the data server 100, the monitor member DB 120, and the information processing device 50 are operated by the research company. Note that the information processing device 50 may be operated by a company other than the research company. The communication network 1 is connected to terminal devices 21, 22, and 23 used by monitor members of the research company. Note that the monitor members are not limited to the example shown in FIG. 1. Terminal devices 11, 12, 13, and 14 used by personnel of companies A to D are also connected to the communication network 1. Companies A to D are companies in different genres A to D, respectively. In the example of FIG. 1, only companies A to D are shown, but this is not limiting. Terminal devices 11 to 14, 21 to 23, and 41 can be configured, for example, as personal computers or tablet terminals.

Genres can be classified into, for example, finance, construction/real estate, logistics/transportation, IT/media, energy/resources, automobiles/machinery, electrical/precision, food, retail/wholesale, lifestyle-related, services, food/drink, entertainment/leisure, etc., but these categories can also be further subdivided.

Companies in each category can request surveys from research companies. Terminal devices 11-14 can send survey requests to terminal device 41. The research company can deliver the survey results to the companies. Terminal device 41 can send the survey results to terminal devices 11-14.

Figure 2 shows an example of a survey response by a monitor member. A survey request is sent from terminal devices 11-14 of companies A-D to terminal device 41. When a survey company representative operates terminal device 41, data server 100 selects an appropriate monitor member in response to the survey request from company A and sends a survey response request to, for example, monitor member terminal devices 21-23. The survey response request includes the survey questions and answer options for each question. When data server 100 receives responses from monitor members' terminal devices 21-23, it can create and update a response result DB for each monitor member. In the example of Figure 2, response result DBs 121, 122, and 123 for monitor members M1, M2, and M3 are recorded in survey DB 110. The same process is performed for survey requests from companies B-D. In this way, by continuously and repeatedly requesting survey responses from monitor members, a response result DB for each monitor member can be built. In this specification, a questionnaire is a document that lists questions and corresponding answers to those questions. Answers may be selectable from multiple options, or may include a blank answer field in which an answer can be written freely. Furthermore, questionnaire responses may include not only the answers themselves, but also the answers and the questions corresponding to those answers.

Next, we will explain how to create a digital twin of the monitor (monitor member). In this specification, a digital twin is a digital twin of a human being, meaning a model that reproduces a person's way of thinking, values, personality, behavioral tendencies, etc., and can operate autonomously in a virtual society just like the monitor himself.

Figure 3 is a diagram showing an example of the configuration of an information processing device 50. The information processing device 50 includes a control unit 51 that controls the entire device, a communication unit 52, a questionnaire analysis unit 53, a compatibility determination unit 54, a questionnaire selection unit 55, a memory unit 56, a tuning unit 57, and a digital twin DB 58.

The control unit 51 can be composed of a CPU (Central Processing Unit), ROM (Read Only Memory), RAM (Random Access Memory), etc. The control unit 51 may also include a GPU (Graphics Processing Unit).

The communication unit 52 includes a communication module and can communicate with terminal devices 11-14, terminal devices 21-23, terminal device 41, survey server 30, and data server 100 via communication network 1.

The control unit 51 can access the survey DB 110 of the data server 100 via the communication unit 52 and read out the response results from the monitor member response result DBs 121-123. In other words, the control unit 51 functions as a collection unit and can collect the survey responses from the monitors themselves.

The storage unit 56 can be configured with a hard disk or semiconductor memory, and stores a general-purpose natural language processing model 561. The general-purpose natural language processing model 561 is a general-purpose model constructed by performing unsupervised pre-training using a large set of sentences. Examples of the general-purpose natural language processing model 561 that can be used include BERT (Bidirectional Encoder Representations from Transformers), GPT (Generative Pre-trained Transformer)-2, GPT-3, and ALBERT (A Lite BERT). Instead of storing the general-purpose natural language processing model 561 in the storage unit 56, the information processing device 50 may access an external language processing server and read it out.

The tuning unit 57 functions as an adjustment unit, and can use the response results collected by the control unit 51 as learning data to tune (adjust) the monitor's digital twin. Here, tuning means generating a digital twin of the monitor based on the general-purpose natural language processing model 561 and updating the generated digital twin, thereby bringing the general-purpose natural language processing model 561 closer to the monitor's digital twin. The tuning unit 57 can be configured with hardware such as a CPU, GPU, DSP (Digital Signal Processor), and FPGA (Field-Programmable Gate Arrays).

The digital twin DB 58 stores the digital twins of each monitor member tuned by the tuning unit 57.

Figure 4 shows an example of creating a digital twin for each monitor member. As shown in Figure 4, a digital twin of monitor M1 can be generated by using the survey questions and answers collected in monitor M1's survey response result DB121 as training data for the general-purpose natural language processing model 561. Similarly, a digital twin of monitor M2 can be generated by using the survey questions and answers collected in monitor M2's survey response result DB122 as training data for the general-purpose natural language processing model 561. Similarly, a digital twin of monitor M3 can be generated by using the survey questions and answers collected in monitor M3's survey response result DB123 as training data for the general-purpose natural language processing model 561. In this way, by performing additional learning on the general-purpose natural language processing model 561 using the monitor's own survey response results, a digital twin incorporating the monitor's own thoughts and values can be generated and updated.

The response result DB contains answers (to each questionnaire question) that are based on the monitor member's way of thinking, values, personality, and behavioral tendencies, and therefore contains a relatively small number of labeled sentences that are sufficient for understanding the way each monitor member thinks, values, personality, and behavioral tendencies. By using the questionnaire questions and answers collected in the response result DB for each monitor member as learning data and performing additional learning on the general-purpose natural language processing model 561, it is possible to adjust the digital twin of each monitor and construct a human digital twin.

Figure 5 is a diagram showing an example of tuning the monitor's digital twin. Assume that a questionnaire is provided to the monitor, and answers A1, A2, A3, ... are obtained to questions Q1, Q2, Q3, .... When the questions included in the questionnaire are input to the monitor's digital twin (which, before additional learning, is the general-purpose natural language processing model 561), the tuning unit 57 can tune the digital twin so that it outputs answers to the questions.

More specifically, the tuning unit 57 can collect both the monitor's own answers and the digital twin's answers to the same questions, and tune the digital twin based on the collected answers. In this case, the monitor's answers can be used as training data to tune the digital twin so that the output data output by the monitor's digital twin matches the monitor's answers. This makes it possible to construct a digital twin that closely resembles the monitor's way of thinking, values, personality, behavioral tendencies, etc.

Next, we will explain the evaluation axes and the evaluation indicators included in the evaluation axes for evaluating how closely the digital twin resembles the monitor (degree of compatibility, degree of growth) in terms of personality and characteristics such as way of thinking, values, individuality, and behavioral tendencies. While any number of axes can be used, this specification uses Spranger's value typology as the first evaluation axis and the ICF (International Classification of Functioning, Disability and Health) as the second evaluation axis. Note that the evaluation axes are not limited to these.

Figure 6 shows the evaluation indicators included in Spranger's value typology (first evaluation axis). Spranger's value typology classifies the areas in which people place the most value in their lives into six types of values (evaluation indicators). The characteristics that form an individual's personality are classified into six types of values, and when people make decisions and take action, they follow one of these six values. As shown in Figure 6, the power-oriented type represents people who value social wins and losses and live for power. The economic-oriented type represents people who value economic gain and loss and love money. The social-oriented type represents people who value cooperation and live for friends and love. The aesthetic-oriented type represents people who value beauty and live by their senses. The theory-oriented type represents people who value logic and love the pursuit of reasoning. The religion-oriented type represents people who value religious morality and love the spiritual world. Note that the content of the value index explanation is not limited to the example in Figure 6, and other similar explanations may be used.

Figure 7 shows an example of a questionnaire regarding the first evaluation axis. The questionnaire shown in Figure 7 contains questions and options from which answers to the questions can be selected. The questionnaire regarding the first evaluation axis contains questions and answers that will yield evaluation values for the six evaluation indicators (six types of values) included in the first evaluation axis.

For example, in a questionnaire regarding power orientation, question Qa1 asks, "When purchasing clothes, I choose based on brand," and there are five answer options, 1 to 5, where 1 is "doesn't apply," 2 is "doesn't really apply," 3 is "can't say," 4 is "somewhat applies," and 5 is "applies." The higher the number in the answer option, the stronger the power orientation. Note that in the example in Figure 7, there are five answer options, but this is not limited to this and three, seven, etc. may also be used. The evaluation value of the evaluation index can be determined based on the number in the answer option. The same applies to other questions (not shown) in the questionnaire regarding power orientation.

In the questionnaire regarding economic orientation, question Qb1 asks, "The higher the salary, the more rewarding my work feels," with five answer options from 1 to 5. The same applies to other questions (not shown). In the questionnaire regarding social orientation, question Qc1 asks, "I like hanging out with a large group of people," with five answer options from 1 to 5. The same applies to other questions (not shown). In the questionnaire regarding aesthetic orientation, question Qd1 asks, "New houses are better than used houses," with five answer options from 1 to 5. The same applies to other questions (not shown). In the questionnaire regarding theoretical orientation, question Qe1 asks, "I plan my trips carefully," with five answer options from 1 to 5. The same applies to other questions (not shown). In the questionnaire regarding religious orientation, question Qf1 asks, "I always visit my ancestors' graves," with five answer options from 1 to 5. The same applies to other questions (not shown).

Figure 8 shows the evaluation indicators included in the International Classification of Functioning, Disability and Health (second evaluation axis). The International Classification of Functioning, Disability and Health (IFSD) considers the functioning model to comprehensively capture the entire human experience. The functioning model is divided into three levels (evaluation indicators): physical and mental functions and structures, activities, and participation. The functioning model provides a common perspective and understanding of all health issues people face in their daily lives, regardless of whether they have a disease or disability. As shown in Figure 8, physical and mental functions and structures are directly linked to maintaining life. Physical functions represent functions such as limb movement, vision and hearing, internal organs, and the mind, while physical structures represent body parts such as the finger joints, stomach, intestines, and skin. Activities represent specific actions that have a purpose in life and consist of a series of movements. Participation represents involvement and fulfillment of roles in various situations in the home and society. The explanation of the value indicators is not limited to the example in Figure 8 and may be other similar explanations.

Figure 9 shows an example of a questionnaire regarding the second evaluation axis. The questionnaire shown in Figure 9 contains questions and options from which answers to the questions can be selected. The questionnaire regarding the second evaluation axis contains questions and answers that will yield evaluation values for the three evaluation indicators (three levels) included in the second evaluation axis.

For example, in a questionnaire regarding physical and mental functions and structure, question Qs1 asks, "I find it difficult to go outside," and there are five answer options, 1 to 5, where 1 is "does not apply," 2 is "does not really apply," 3 is "can't say," 4 is "somewhat applies," and 5 is "applies." The higher the number in the answer option, the worse the health condition. Note that in the example in Figure 9, there are five answer options, but this is not limited to this and two options (yes, no, etc.) are also possible. The evaluation value of the evaluation index can be determined based on the number in the answer option. The same applies to other questions (not shown) in the questionnaire regarding physical and mental functions and structure.

In the survey regarding activities, question Qt1 states, "I go out every day to work, school, or shopping," with five answer choices from 1 to 5. The same applies to other questions (not shown). In the survey regarding participation, question Qu1 states, "I participate in volunteer or community activities," with five answer choices from 1 to 5. The same applies to other questions (not shown).

The questionnaire analysis unit 53 can analyze the response results of the questionnaire output by the digital twin. More specifically, the questionnaire analysis unit 53 can calculate an evaluation value for each evaluation index by aggregating the answer options for each evaluation index from the digital twin. For example, the answer options for each of questions Q1 to Qn for a certain evaluation index are A1 to An. A1 to An are, for example, any number from 1 to 5. The evaluation value EE of a certain evaluation index can be calculated using the formula EE = (A1 + A2 + ... + An) / n. The questionnaire analysis unit 53 may also be equipped with language processing functions. For example, if responses to a questionnaire are not selected by selecting an option but are instead freely entered, the questionnaire analysis unit 53 may read the text entered in the response field and perform processing such as morphological analysis, syntactic analysis, semantic analysis, and contextual analysis to calculate an evaluation value for each evaluation index.

Next, we will explain the degree of conformance, which indicates how closely the digital twin resembles the monitor. Below, we will explain the degree of conformance in the first evaluation axis and the degree of conformance in the evaluation indicators included in the first evaluation axis, but the same applies to the second evaluation axis.

Figure 10 shows the degree of conformance of the digital twin on the evaluation axes. In Figure 10, the six evaluation indicators of the first evaluation axis, a: power orientation, b: economic orientation, c: social orientation, d: aesthetic orientation, e: theoretical orientation, and f: religious orientation, are shown in a radar chart. The monitor's own evaluation values for power orientation, economic orientation, social orientation, aesthetic orientation, theoretical orientation, and religious orientation are Ma, Mb, Mc, Md, Me, and Mf, respectively, and the digital twin's evaluation values are Da, Db, Dc, Dd, De, and Df.

The evaluation value ME of the monitor's own evaluation axis can be expressed by the formula ME = (Ma + Mb + Mc + Md + Me + Mf) / 6. In other words, the evaluation value of the evaluation axis can be expressed as the average value of the evaluation values of the evaluation indicators included in the evaluation axis. Similarly, the evaluation value DE of the digital twin's evaluation axis can be expressed by the formula DE = (Da + Db + Dc + Dd + De + Df) / 6.

The fitness GF (%) of a digital twin can be expressed by the formula GF = {1 - (|ME-DE|)/ME} x 100. If the indicators used to evaluate the degree of growth of a digital twin are, for example, "mature," "growing," and "apprentice," then "mature" is when the fitness GF is 80% or higher, "growing" is when the fitness GF is 50% to 80%, and "apprentice" is when the fitness GF is less than 50%. Note that the fitness and digital twin evaluation are merely examples and are not limited to the example in Figure 10. Furthermore, if the fitness on the first evaluation axis is GF1 and the fitness on the second evaluation axis is GF2, the overall fitness GFT can be calculated using the formula GF1 x w1 + GF2 x w2. Here, w1 and w2 are weighting coefficients. If w1 = w2 = 0.5, the overall fitness GFT is the average value of the fitness on each evaluation axis. The weighting coefficients w1 and w2 can be set appropriately depending on the characteristics of the evaluation axis, etc.

Figure 11 shows the degree of conformance for digital twin evaluation indices. As shown in Figure 11, the monitor's own evaluation values for power orientation, economic orientation, social orientation, aesthetic orientation, theoretical orientation, and religious orientation are Ma, Mb, Mc, Md, Me, and Mf, respectively, and the digital twin's evaluation values are Da, Db, Dc, Dd, De, and Df. The conformance GFi (%) for each digital twin evaluation indices can be expressed by the formula GFi = {1 - (|Mi - Di|)/Mi} x 100, where i is any of a to f.

As described above, the compatibility determination unit 54 can determine the compatibility of a digital twin. The compatibility may be determined using an evaluation axis, or using evaluation indicators included in the evaluation axis. Furthermore, the compatibility determination unit 54 can determine the evaluation of the digital twin ("mature," "growing," or "apprentice") as shown in FIG. 10 depending on the compatibility. The compatibility determined by the compatibility determination unit 54 may be output via the communication unit 52 to, for example, the questionnaire server 30 or the terminal device 41.

Furthermore, even if the monitor himself answers the same questionnaire, the answer may not always be the same; the answer may vary depending on the time or place of answering the questionnaire. Therefore, when calculating the evaluation value of an evaluation axis or evaluation index, the monitor may be asked to answer the same questionnaire a predetermined number of times, and the average of the evaluation values for each answer may be used as the final evaluation value.

If the compatibility determination unit 54 determines that the digital twin is, for example, "growing" or "apprentice," the tuning unit 57 can collect further questionnaire responses from the monitor and use the collected responses as learning data to tune the monitor's digital twin. This allows the digital twin to grow further and become closer to the monitor's way of thinking, values, personality, or behavioral tendencies.

Next, we'll explain how to select a survey.

Figure 12 shows an example of a questionnaire classification table. Figure 12 shows a collection of many questionnaires categorized by two aspects: evaluation axis (evaluation index) and genre (company). For example, when viewed from the perspective of company (genre), it can be seen that the questionnaire group related to company A also includes questionnaire groups related to power orientation, social orientation, theory orientation, and mental and physical functions/physical structure. On the other hand, when viewed from the perspective of evaluation index, it can be seen that the questionnaire group related to power orientation also includes questionnaire groups related to companies A, D, E, and F.

Furthermore, the questionnaire group regarding Company E also includes a questionnaire group regarding the evaluation indexes for the first evaluation axis and the evaluation indexes for the second evaluation axis, so it can be seen that by collecting questionnaire responses using the questionnaire group regarding Company E, responses regarding the evaluation indexes for the first evaluation axis and the evaluation indexes for the second evaluation axis can be obtained.

The questionnaire selection unit 55 can select questionnaires to collect to further tune the digital twin based on the degree of conformance (evaluation results) for the evaluation axis determined by the conformance determination unit 54, or the degree of conformance (evaluation results) for the evaluation indicators included in the evaluation axis.

For example, if the digital twin's compatibility with the monitor in terms of economic orientation is low, a questionnaire can be selected from the group of questionnaires related to companies B, E, or F, and the selected questionnaire can be used to tune the monitor's digital twin in order to improve compatibility with the monitor in terms of economic orientation.

Furthermore, if the digital twin's compatibility with the monitor in terms of power orientation, social orientation, and theory orientation is low, a questionnaire can be selected from the questionnaire group related to Company A and used to tune the monitor's digital twin in order to improve compatibility with the monitor in each evaluation index.

Furthermore, if the overall compatibility of the digital twin is low, in order to increase the overall compatibility with the monitor, a questionnaire can be selected from the questionnaire group related to Company E, which includes a group of questionnaires related to all evaluation indicators, and the selected questionnaire can be used to tune the monitor's digital twin.

As described above, the questionnaire selection unit 55 selects questionnaires to collect so that they include responses regarding evaluations for each of the required number of evaluation axes for evaluating the compatibility between the digital twin and the monitor, and can use the collected questionnaires to tune the monitor's digital twin. The selection of questionnaires by the questionnaire selection unit 55 may be performed when additional learning is performed for the first time on the general-purpose natural language processing model 561, or may be performed when further tuning the digital twin based on the evaluation results of the evaluation indicators.

The survey selection unit 55 may also select surveys to collect based on the monitor's own recommendation. This allows the monitor to select surveys based on their own judgment of their way of thinking, values, personality, and behavioral tendencies.

Figure 13 is a diagram showing the processing steps for tuning a digital twin by the information processing device 50. For convenience, the following description will be given with the control unit 51 as the main actor in the processing. The control unit 51 collects questionnaire responses from the monitors themselves (S11) and uses the collected responses as learning data to tune the monitors' digital twins (S12). An example of tuning a digital twin is shown in Figure 5.

The control unit 51 calculates the evaluation value of the evaluation index for each evaluation axis (S13), calculates the degree of fit for each evaluation index (S14), and calculates the degree of fit for each evaluation axis and the overall degree of fit (S15). Examples of the degree of fit for each evaluation index are shown in Figure 11. Examples of the degree of fit for each evaluation axis are shown in Figure 10. The overall degree of fit can be the average value of the degrees of fit for each evaluation axis.

The control unit 51 determines whether the state of the digital twin is mature (S16). Whether the digital twin is mature may be determined, for example, by whether the conformance of each of the first evaluation axis and the second evaluation axis is 80% or higher, or by whether the overall conformance is 80% or higher. If the state of the digital twin is not mature (NO in S16), the control unit 51 determines (selects) the evaluation indexes for the questionnaire to be collected based on the evaluation results of the evaluation axes or the evaluation indexes included in the evaluation axes (S17).

The control unit 51 determines whether or not the monitor has made a recommendation (S18), and if so (YES in S18), selects a questionnaire to collect based on the monitor's recommendation (S19), and continues processing from step S11 onwards. If the monitor has not made a recommendation (NO in S18), the control unit 51 continues processing from step S11 onwards. If the digital twin is in a mature state (YES in S16), the control unit 51 ends processing.

The information processing device 50 can also be realized using a general-purpose computer equipped with a CPU (processor), ROM, RAM (memory), etc. In other words, the information processing device 50 can be realized on a computer by loading a computer program, such as that shown in FIG. 13, that defines the procedures for each process into the computer's RAM (memory) and executing the computer program on the CPU (processor). The computer program may be installed on a recording medium or may be installed on the computer via a network.

As described above, according to this embodiment, it is possible to create a digital twin that incorporates the individual's way of thinking and criteria for judgment by taking the opportunity to answer a questionnaire. Furthermore, for information that is lacking in order to build knowledge equivalent to that of the individual, the digital twin can select a questionnaire and answer it, thereby bringing the individual's way of thinking and criteria for judgment even closer to that of the individual.

Second Embodiment
In the second embodiment, a questionnaire system and a questionnaire method using the constructed digital twin will be described.

Figure 14 shows an example of the configuration of the survey server 30. The survey server 30 includes a control unit 31 that controls the entire server, a communication unit 32, a selection unit 33, a selection unit 34, an identification unit 35, a digital twin DB 36, a tuning unit 37, a memory unit 38, and a survey DB 39.

The control unit 31 can be composed of a CPU (Central Processing Unit), ROM (Read Only Memory), RAM (Random Access Memory), etc. The control unit 31 may also include a GPU (Graphics Processing Unit).

The communication unit 32 includes a communication module and can communicate with terminal devices 11-14, terminal devices 21-23, information processing device 50, and data server 100 via communication network 1.

The memory unit 38 can be configured as a hard disk or semiconductor memory, etc., and stores the required information.

The digital twin DB 36 stores all or part of the digital twins stored in the digital twin DB 58 of the information processing device 50. In other words, the survey server 30 has multiple digital twins of multiple monitor members generated using the survey responses of multiple monitor members. Note that instead of being configured to be provided in the survey server 30, the digital twin DB 36 may be an external DB.

The control unit 31 controls multiple digital twins to respond to the questionnaire.

Figure 15 shows an example of a survey method using the survey server 30. The survey server 30 realizes a virtual space (cyberspace) that is distinct from real space, and can perform calculations (computations) that enable digital twins to operate in the virtual space. As shown in Figure 15, survey requests are sent from terminal devices 11-14 of companies A-D to terminal device 41. When a survey company representative operates terminal device 41, the survey server 30 outputs a survey to the digital twins corresponding to each monitor member, causing the digital twins to respond to the survey. The survey server 30 obtains the responses from the digital twins and transmits survey results based on the responses to terminal devices 11-14 of companies A-D.

As mentioned above, by having the digital twins corresponding to the monitor members answer the survey instead of the monitor members, the survey results can be provided quickly (almost in real time compared to survey results based on the responses of the monitor members), and survey responses can be collected in a timely manner. Furthermore, by using a large number of digital twins, large-scale surveys can be conducted. Furthermore, even if the survey has a large number of questions or the questions themselves are long, digital twins can answer them instantly, preventing surveys from going unanswered. Furthermore, by eliminating the need to send or mail surveys to monitor members, surveys can be conducted inexpensively.

Figure 16 is a diagram showing an example of the configuration of the digital twin DB 36. The digital twin DB 36 records the characteristics of digital twins in association with the ID (identification code) of the digital twin tuned by the information processing device 50. As shown in Figure 16, the digital twin DB 36 includes the following components: digital twin ID, monitor member ID, overall compatibility GFT, compatibility on the first evaluation axis GF1, compatibility on the second evaluation axis GF2, genre, and attributes. The monitor member ID is the ID of the monitor member reproduced by the digital twin. The overall compatibility GFT, compatibility on the first evaluation axis GF1, and compatibility on the second evaluation axis GF2 each represent the compatibility of the digital twin with the monitor member. The genre includes one or more genres and represents the genres that the digital twin can handle. For example, it is possible to identify digital twins that can respond to a questionnaire about the genre of a survey from a company. The attributes represent the attributes of the monitor member reproduced by the digital twin.

The terminal device 41 can access the digital twin DB 36 and display each configuration of the digital twin DB 36. In other words, the control unit 31 can output the compatibility between each of the multiple digital twins and the monitor member.

Next, we will explain how to design a survey when requested by a company.

Figure 17 is a diagram showing an example of the survey design screen 210. The survey design screen 210 can be displayed on the terminal device 41. On the survey design screen 210, parameters can be selected for each monitor member and digital twin. For monitor members, the number of people participating in the survey, attributes, and related genres can be selected. The number of people can be 0. If a number of people of 0 is selected, only the digital twin will be able to respond to the survey. Attributes include, for example, male, female, unmarried, married, age (generation), etc. However, attributes are not limited to these. One or multiple genres can be selected.

For digital twins, you can select the number of people to respond to the survey, their attributes, related genres, and compatibility. By selecting a number other than 0 for the number of monitor members and digital twins, you can have both monitor members and digital twins respond to the survey.

The selection unit 34 can select respondents to answer the questionnaire from both monitor members and digital twins. This allows respondents to be selected according to the company's needs.

The compatibility level can be selected within a range between a lower limit and an upper limit. If only the lower limit is selected, digital twins with a compatibility level equal to or greater than the lower limit can be selected. In other words, the selection unit 33 can select digital twins to respond to the questionnaire based on the selected compatibility level.

By operating the "Run" icon, you can design a survey. Specifically, you can design a survey based on the selected parameters. By operating the "Clear" icon, you can select the parameters again.

Figure 18 is a diagram showing an example of the configuration of the questionnaire DB39. The questionnaire DB39 stores a group of questionnaires (consisting of multiple questionnaires) that have been previously classified and collected in association with genres and attributes. Attributes include, for example, male, female, unmarried, married, and age groups such as 20s to 40s and 40s to 60s. For example, if the genre required for answering a questionnaire is A and the respondent attributes are male and 20s to 40s, then a questionnaire is designed from questionnaire groups QA11 and QA15.

In other words, when the control unit 31 receives a survey request from a company's terminal devices 11-14 via the communication unit 32, the identification unit 35 can identify the genre of the survey and the attributes of the respondent. The control unit 31 can cause multiple digital twins selected based on the genre and attributes identified by the identification unit 35 to respond to the survey.

The same applies to other genres B to D and attributes. This makes it easy to design an appropriate survey depending on the survey genre and the attributes of the respondents. The survey may be designed by a research company, or may be designed and created by each company. In this case, the survey server 30 can obtain surveys from the company's terminal devices 11 to 14 and have the digital twins and monitor members respond to the surveys.

Figure 19 is a diagram showing an example of the monitor member response screen 220. The monitor member response screen 220 can be displayed on the terminal device 41. When only monitor members are allowed to respond to the survey, the monitor member response screen 220 displays a list of members who have responded and a list of members who have not responded. As shown in Figure 19, the monitor member response list displays the member IDs of members who have responded to the survey and their response results.

The list of unanswered monitor members displays the member IDs of members who have not answered the survey, and whether or not they have given permission for their digital twins to answer on their behalf. Permission indicates that the monitor member has given permission for their digital twin to answer on their behalf. Permission not indicates that permission has not been granted, or there is no response regarding permission. If you select "Permission only" for the alternative answer and operate the "Execute digital twin answer" icon, the survey will be output to the digital twins corresponding to members who have not answered but have given permission, and you will be able to obtain alternative answers using digital twins. If you select "All" for the alternative answer and operate the "Execute digital twin answer" icon, the survey will be output to the digital twins corresponding to members who have not answered, and you will be able to obtain alternative answers using digital twins. Permission for alternative answers can also be decided when registering as a monitor member.

As described above, the control unit 31 can send questionnaires to the monitor members' terminal devices 21-23 via the communication unit 32 and receive questionnaire responses from the terminal devices 21-23. If a monitor member has not responded, the control unit 31 can have the digital twin corresponding to that monitor member respond to the questionnaire. This allows the digital twin to respond on behalf of the monitor member even if the monitor member alone is unable to provide sufficient responses, ensuring that the questionnaire survey can be conducted reliably.

Furthermore, the control unit 31 can have the digital twin corresponding to the monitor member answer a questionnaire in response to a request from the monitor member. This can improve convenience for the monitor member.

The tuning unit 37 has the same functions as the tuning unit 57 of the information processing device 50. The tuning unit 37 may tune a digital twin by having a digital twin whose compatibility with the monitor member is not within the required range answer a questionnaire. This makes it possible to realize a questionnaire system that can develop digital twins.

Figure 20 is a diagram showing an example of the processing steps of a survey method by the survey server 30. For convenience, the following description will be given with the control unit 31 as the main processor. The control unit 31 acquires a survey request (S31) and identifies the survey genre and the attributes of the respondent (S32). The control unit 31 determines whether responses are required from both the monitor member and the digital twin (S33). Whether responses are required from both the monitor member and the digital twin can be determined, for example, based on the survey request acquired from the company.

If responses are not required from both (NO in S33), the control unit 31 inputs the questionnaire into the digital twin (S34), obtains the questionnaire responses from the digital twin (S35), and performs the processing of step S45 described below. If responses are required from both the monitor members and the digital twin (YES in S33), the control unit 31 selects the number of respondents from the monitor members and the digital twin (S36). If the number of respondents from both parties is specified in the questionnaire survey request, that number should be followed. If the number of respondents from both parties is not specified in the questionnaire survey request, the number should be determined appropriately depending on the questionnaire genre and the attributes of the respondents.

The control unit 31 sends a questionnaire to the monitor members' terminal devices 21-23 (S37) and inputs the questionnaire into the digital twin (S38). The control unit 31 acquires the questionnaire responses from the digital twin (S39) and receives the questionnaire responses from the monitor members' terminal devices 21-23 (S40). Note that responses from the digital twin can be acquired almost in real time, but responses from the monitor members may take several days or even a week.

The control unit 31 determines whether or not the monitor member has left any unanswered responses (S41), and if there are no unanswered responses (NO in S41), performs the processing of step S45 described below. If there are unanswered responses (YES in S41), the control unit 31 determines whether or not to provide an alternative response using the digital twin (S42). If an alternative response is to be provided (YES in S42), the control unit inputs the questionnaire into the digital twin (S43), obtains the questionnaire responses from the digital twin (S44), and performs the processing of step S45 described below. If an alternative response is not to be provided (NO in S42), the control unit 31 generates survey results (S45) and ends the processing.

The survey server 30 can also be implemented using a general-purpose computer equipped with a CPU (processor), ROM, RAM (memory), etc. In other words, a computer program that defines the procedures for each process, such as that shown in FIG. 20, can be loaded into the computer's RAM (memory) and executed by the CPU (processor), thereby implementing the information processing device 50 on the computer. The computer program may be installed on a recording medium or may be installed on the computer via a network.

The information processing method of this embodiment collects questionnaire responses from the monitors themselves, and uses the collected responses as learning data to adjust the monitors' digital twins.

The information processing method of this embodiment adjusts the digital twin so that, when a question included in the questionnaire is entered, an answer to the question is output.

The information processing method of this embodiment collects responses from the monitor and the digital twin to the same questions, and adjusts the digital twin based on the collected responses.

The information processing method of this embodiment determines the evaluation results for each of the required number of evaluation axes for assessing the compatibility between the digital twin and the monitor based on the response, and adjusts the digital twin based on the determined evaluation results.

The information processing method of this embodiment determines the conformance of the digital twin for each evaluation axis based on the determined evaluation results, and outputs the determined conformance.

The information processing method of this embodiment determines the conformance of the digital twin for each of the multiple evaluation indicators included in the evaluation axis based on the response, and outputs the determined conformance for each of the evaluation indicators.

The information processing method of this embodiment selects the questionnaires to collect based on the evaluation results determined for each of the evaluation axes, and uses the selected questionnaires to adjust the digital twin of the monitor.

The information processing method of this embodiment collects a questionnaire that includes responses regarding evaluations for each of the required number of evaluation axes for evaluating the compatibility between the digital twin and the monitor, and uses the collected questionnaire to adjust the monitor's digital twin.

The information processing method of this embodiment selects surveys to collect based on the monitor's own recommendations, and uses the selected surveys to adjust the monitor's digital twin.

The information processing device of this embodiment includes a collection unit that collects questionnaire responses from the monitors themselves, and an adjustment unit that uses the collected responses as learning data to adjust the monitors' digital twins.

The computer program of this embodiment causes a computer to collect questionnaire responses from the monitor himself/herself and use the collected responses as learning data to adjust the monitor's digital twin.

The survey system of this embodiment includes a control unit that controls a digital twin corresponding to a monitor member, generated using the monitor member's responses to the survey, to respond to the survey.

The questionnaire system of this embodiment includes an output unit that outputs the degree of compatibility between the monitor member and the digital twin corresponding to the monitor member, and a selection unit that selects the digital twin to respond to the questionnaire based on the degree of compatibility.

The questionnaire system of this embodiment includes a selection unit that selects respondents to answer the questionnaire from both the monitor members and the digital twins.

The survey system of this embodiment includes an adjustment unit that adjusts a digital twin by having the digital twin, whose compatibility with the monitor member is not within the required range, respond to the survey.

In the survey system of this embodiment, the control unit causes the digital twin corresponding to the monitor member to answer the survey in response to a request from the monitor member.

The survey system of this embodiment includes a sending unit that sends surveys to multiple monitor members and a receiving unit that receives responses to the survey, and if a monitor member has not responded, the control unit causes the digital twin corresponding to that monitor member to respond to the survey.

The questionnaire system of this embodiment includes an acquisition unit that acquires questionnaire survey requests from company terminal devices, an identification unit that identifies the genre of the questionnaire and the attributes of the respondents, and multiple digital twins classified based on the genre and attributes, and the control unit causes multiple digital twins selected based on the identified genre and attributes to respond to the questionnaire.

The computer program of this embodiment causes a computer to execute a process that causes a digital twin corresponding to a monitor member, generated using the monitor member's responses to a survey, to respond to the survey.

The survey method of this embodiment acquires a survey, accepts the selection of the number of respondents, and has digital twins corresponding to the selected number of monitor members, generated using the survey responses from monitor members, respond to the acquired survey.

REFERENCE SIGNS LIST 1 Communication network 11, 12, 13, 14, 21, 22, 23, 41 Terminal device 30 Questionnaire server 31 Control unit 32 Communication unit 33 Selection unit 34 Selection unit 35 Identification unit 36 Digital twin DB
37 Tuning section 38 Memory section 39 Questionnaire DB
50 Information processing device 51 Control unit 52 Communication unit 53 Questionnaire analysis unit 54 Goodness of fit determination unit 55 Questionnaire selection unit 56 Storage unit 561 General-purpose natural language processing model 57 Tuning unit 58 Digital twin DB
100 Data server 110 Survey DB
121, 122, 123 Answer result DB
120 Monitor Member DB

Claims (10)

The control unit collects the questionnaires completed by the monitors themselves from the response result database for each monitor ,
The control unit generates or updates a digital twin of the monitor using the questionnaire collected in the response result DB as learning data for a general -purpose natural language processing model,
The control unit collects the responses of the monitor and the digital twin to the same questions in the questionnaire ,
The control unit generates or updates the digital twin based on the collected answers,
Based on the response, the control unit determines the evaluation results for each of the required number of evaluation axes for evaluating the compatibility between the digital twin and the monitor,
The control unit selects a questionnaire to be collected based on the evaluation results for each of the evaluation axes determined,
The control unit updates the monitor's digital twin using the selected questionnaire .
Information processing methods. The control unit collects the questionnaires completed by the monitors themselves from the response result database for each monitor ,
The control unit generates or updates a digital twin of the monitor using the questionnaire collected in the response result DB as learning data for a general -purpose natural language processing model,
The control unit selects questionnaires to be collected based on the evaluation results for each of the required number of evaluation axes for evaluating the compatibility between the digital twin and the monitor,
The control unit updates the monitor's digital twin using the selected questionnaire.
Information processing methods. The control unit collects the questionnaires completed by the monitors themselves from the response result database for each monitor ,
The control unit generates or updates a digital twin of the monitor using the questionnaire collected in the response result DB as learning data for a general-purpose natural language processing model,
The control unit determines whether or not the monitor himself/herself has recommended the product,
If there is a recommendation, the control unit selects a questionnaire to collect;
The control unit updates the monitor's digital twin using the selected questionnaire.
Information processing methods. When a question included in the questionnaire is entered, the control unit generates or updates the digital twin so as to output an answer to the question.
4. The information processing method according to claim 1. The control unit determines the conformance of the digital twin for each evaluation axis based on the determined evaluation result,
The control unit outputs the determined degree of conformance.
The information processing method according to claim 1 . The control unit determines the conformance of the digital twin for each of a plurality of evaluation indexes included in the evaluation axis based on the responses to the questionnaire ,
The control unit outputs the determined degree of conformance for each of the evaluation indexes.
6. The information processing method according to claim 1 or claim 5. A control unit is provided,
The control unit
We collect the questionnaires from the monitors themselves from the DB of response results for each monitor ,
Using the questionnaire collected in the response result DB as learning data for a general-purpose natural language processing model, a digital twin of the monitor is generated or updated;
Collect the responses of the monitor and the digital twin to the same questions in the questionnaire ,
generating or updating the digital twin based on the collected responses;
Based on the responses, an evaluation result is determined for each of the required number of evaluation axes for evaluating the compatibility between the digital twin and the monitor;
Selecting questionnaires to be collected based on the evaluation results for each of the evaluation axes determined;
updating said participant's digital twin with the selected survey ;
Information processing device. A control unit is provided,
The control unit
We collect the questionnaires from the monitors themselves from the DB of response results for each monitor ,
Using the questionnaire collected in the response result DB as learning data for a general-purpose natural language processing model, a digital twin of the monitor is generated or updated;
Selecting questionnaires to be collected based on the evaluation results for each of the required number of evaluation axes for evaluating the compatibility between the digital twin and the monitor;
updating said participant's digital twin with the selected survey;
Information processing device. On the computer,
We collect the questionnaires from the monitors themselves from the DB of response results for each monitor ,
Using the questionnaire collected in the response result DB as learning data for a general-purpose natural language processing model, a digital twin of the monitor is generated or updated;
Collect the responses of the monitor and the digital twin to the same questions in the questionnaire ,
generating or updating the digital twin based on the collected responses;
Based on the responses, an evaluation result is determined for each of the required number of evaluation axes for evaluating the compatibility between the digital twin and the monitor;
Selecting questionnaires to be collected based on the evaluation results for each of the evaluation axes determined;
updating said participant's digital twin with the selected survey ;
A computer program that executes a process. On the computer,
We collect the questionnaires from the monitors themselves from the DB of response results for each monitor ,
Using the questionnaire collected in the response result DB as learning data for a general-purpose natural language processing model, a digital twin of the monitor is generated or updated;
Selecting questionnaires to be collected based on the evaluation results for each of the required number of evaluation axes for evaluating the compatibility between the digital twin and the monitor;
updating said participant's digital twin with the selected survey;
A computer program that executes a process.