Learning model generation method, challenge support system and program

Abstract

To provide a learning model generation device for generating a learning model capable of presenting a challenge environment in which challenge by a user is highly likely to be achieved.SOLUTION: A learning model generation device includes: an attribute information reference unit 203 for referring to attribute information indicating an attribute of a user having challenge to be achieved; an environmental conditions generating unit 204 for generating challenge environment conditions as environmental conditions that are a set of a plurality of environmental elements stored in a storage unit concerning an environment for performing the challenge on the basis of the attribute information; an achievement level reference unit 208 for referring to a degree of achievement of challenge when the challenge is executed by a user in the challenge environment conditions; a learning data storage unit 209 for storing a set of the degree of achievement, the challenge environment conditions, and the attribute information as learning data in a storage unit 202; and a learning model generation unit 210 for generating a learning model M1 to be learned so that attribute information is input and challenge environment conditions are pieces of output on the basis of the learning data.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to a learning model generation method, a challenge support system, and a program. [Background technology]

[0002] People often have challenges that they want to achieve but are unable to achieve. Challenges can include, for example, walking to maintain or improve health, going for a health check, quitting smoking, studying English conversation, etc.

[0003] One of these types of challenges is the challenge to improve exercise habits for the purpose of maintaining and improving health. Patent Document 1 shows an exercise management device that enables real-time health management and exercise planning that takes into account the impact on health, and enables users to easily perform efficient exercise. The exercise management device described in Patent Document 1 manages the user's exercise using an analysis model for deriving exercise recommendation information that recommends exercises suitable for the user. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent No. 5694206 Summary of the Invention [Problem to be solved by the invention]

[0005] The specific challenge content of an exercise plan, such as exercise information including duration, number of steps, and speed in the exercise management device described in Patent Document 1, can affect whether or not the user who takes on the challenge can achieve the plan.

[0006] Factors related to a user's achievement of a challenge include not only the challenge to be performed, but also factors due to the external environment in which the user performs the challenge. The external environment is a factor that affects the motivation of the user to perform the challenge. In other words, even if the content of the challenge presented to the user using a certain learning model is itself executable by the user, if the user performs the challenge in an environment where the user lacks motivation to perform the challenge, the execution may be delayed, and as a result, the challenge may not be completed.

[0007] Therefore, an object of the present invention is to provide a learning model generation device that generates a learning model that can present a challenge environment in which a user is likely to complete a challenge. [Means for solving the problem]

[0008] A learning model generation device according to one embodiment of the present invention includes an attribute information reference unit that references attribute information indicating the attributes of a user who has a challenge to be completed, an environmental condition generation unit that generates challenge environmental conditions as environmental conditions, which are a set of multiple environmental elements stored in a memory unit regarding the environment in which the challenge is performed, based on the attribute information, an achievement level reference unit that references the achievement level of the challenge when the challenge is performed by the user under the challenge environmental conditions, a learning data storage unit that stores a set of the achievement level, challenge environmental conditions, and attribute information in the memory unit as learning data, and a learning model generation unit that generates a learning model based on the learning data that is trained to use the attribute information as input and the challenge environmental conditions and the achievement level (estimated value) as output. Effect of the Invention

[0009] According to the present invention, it is possible to provide a learning model generation device that generates a learning model capable of presenting a challenge environment in which a user is likely to complete a challenge. [Brief description of the drawings]

[0010] [Figure 1] 1 is a schematic diagram of a challenge support system according to an embodiment of the present invention. [Diagram 2] FIG. 1 is a block diagram of a learning model generating device according to an embodiment of the present invention. [Diagram 3] FIG. 13 is a diagram illustrating an example of an attribute element according to the embodiment. [Figure 4] FIG. 1 is a diagram illustrating an example of an environmental element according to an embodiment of the present invention. [Diagram 5] FIG. 1 is a diagram illustrating an example of an environmental element according to an embodiment of the present invention. [Figure 6] FIG. 2 is a block diagram of a challenge management device according to the present embodiment. [Figure 7] 13 is an example of challenge information according to the embodiment. [Figure 8] 1 is a diagram illustrating an outline of the processing in the challenge support system according to this embodiment. FIG. [Figure 9] 2 is a diagram illustrating the generation of a first learning model in the learning model generating device according to the present embodiment. FIG. [Figure 10] FIG. 13 is a diagram explaining the generation of a second learning model based on achievement level. [Figure 11] FIG. 13 is a diagram explaining the generation of a second learning model based on acceptance. [Figure 12] A diagram explaining the generation of a second learning model based on input from a user. [Figure 13] 10 is a diagram illustrating an example of challenge management information in a challenge management device. FIG. [Figure 14] 1 is a diagram for explaining an outline of the process of updating the first learning model according to the present embodiment. FIG. [Figure 15] 11 is a diagram illustrating another example of challenge management information in the challenge management device. FIG. [Figure 16] 11A to 11C are diagrams illustrating a display process of an estimated achievement level based on attribute information according to the embodiment. [Figure 17] FIG. 11 is a diagram illustrating an example of attribute information. [Figure 18] FIG. 13 is a diagram illustrating an example of a challenge environmental condition. [Figure 19] FIG. 11 is a diagram illustrating an example of attribute information. [Figure 20]FIG. 13 is a diagram illustrating an example of a challenge environmental condition. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will now be described with reference to the accompanying drawings, in which like reference numerals denote like or similar structures.

[0012] 1 shows an overview of a challenge support system 10 according to this embodiment. The challenge support system 10 includes a learning model generation device 101, user terminals 102a and 102b, and a challenge management device 103. When there is no particular need to distinguish between them, the user terminals 102a and 102b are referred to as user terminals 102. The challenge management device 103 is connected to the learning model generation device 101 and the user terminal 102 via a network N so as to be able to communicate with each other.

[0013] The learning model generation device 101 is an information processing device that executes a predetermined program to generate a learning model that generates environmental conditions for a challenge managed in the challenge management device 103. The learning model generation device 101 is, for example, a server. The environmental conditions and the generation process of the learning model by the learning model generation device 101 will be described later.

[0014] The user terminal 102 transmits and receives necessary information regarding the challenge to and from the challenge management device 103 by executing a predetermined program. The user terminals 102a and 102b each include a communication unit 1021a, 1021b and a display unit 1022a, 1022b. The communication unit 1021 communicates with the challenge management device 103 through a network N. The display unit 1022 displays information acquired from the challenge management device 103 or a screen required for inputting information to be transmitted to the challenge management device 103. The user terminal 102 is, for example, an information processing terminal such as a smartphone, a PC, or a tablet, but is not limited to these. The user terminal 102 is, for example, a smartphone or a tablet, and transmits and receives information to and from the challenge management device 103 by a program installed as an application.

[0015] The challenge management device 103 manages the challenges that the user takes on by executing a predetermined program. The challenge management device 103 manages the challenges by associating the challenge environment conditions of the user's challenge with the user's challenge based on the learning model generated by the learning model generation device 101 and the user's attribute information.

[0016] The challenge managed by the challenge management device 103 is a challenge that is set so that a user deposits his / her own funds on a blockchain network and the funds are distributed depending on whether the challenge is achieved or not, as described in, for example, Japanese Patent No. 6762647. Alternatively, the challenge may be such that a company or the like provides funds to employees and the funds are distributed depending on the challenge results.

[0017] Each unit of the learning model generation device 101 will be described with reference to Fig. 2 and Fig. 3 to Fig. 5. Fig. 2 shows a block diagram of the learning model generation device 101. The learning model generation device 101 has a communication unit 201, a storage unit 202, an attribute information reference unit 203, an environmental condition generation unit 204, an environmental candidate generation unit 205, an environmental element presentation unit 206, a compatibility acquisition unit 207, an achievement level reference unit 208, a learning data storage unit 209, and a learning model generation unit 210. Each unit of the learning model generation device 101 can be realized, for example, by using a storage area such as a memory in the learning model generation device 101, or by a processor executing a program stored in the storage area.

[0018] The communication unit 201 controls communication between the learning model generation device 101 and external terminals and information processing devices via a network N.

[0019] The storage unit 202 stores various types of information used in the processing in the learning model generation device 101. The storage unit 202 includes a learning model storage unit 2021, an attribute element storage unit 2022, an environmental element storage unit 2023, and a compatibility storage unit 2024.

[0020] The learning model storage unit 2021 stores the learning model M1 and the learning model M2 generated by the learning model generation device 101. The learning models M1 and M2 will be described later.

[0021] The attribute element storage unit 2022 stores attribute items, which are items in attribute information indicating the attributes of a user. An example of an attribute item stored in the attribute element storage unit 2022 is shown in Fig. 3. In Fig. 3, attribute items A include items such as "basic attributes," "health," "personality," "values," and "cognitive bias," and attribute items B are provided as sub-items for each attribute item A.

[0022] By storing information about each user as a value corresponding to each attribute item B, attribute information about a certain user is stored in the learning model generating device 101.

[0023] For example, the "cognitive bias" category has the following items as attribute item B: "time discounting," "probability weighting," "risk and loss aversion," "conformity tendency," and "sunk cost." The "time discounting" category is an item about the tendency to underestimate the future value over time compared to the present value when evaluating the value of the reward for a certain action. The "probability weighting" category is an item about the tendency to overestimate events that occur with a low probability and underestimate events that occur with a high probability. The "risk and loss aversion" category is an item about the tendency to make decisions that avoid risky challenges and to make decisions that place more importance on losses than on gains. The "conformity tendency" category is an item about the tendency to be influenced by the opinions and actions of people around you and to act in the same way. The "sunk cost" category is an item about the tendency to seek psychological justification for costs that are already impossible to recover. In other words, cognitive bias here refers to a psychological disposition that affects the execution of some action or judgment on something.

[0024] When the user performs a challenge, environmental elements, which are items in the environmental conditions that indicate the user's external environment, are stored in the environmental element storage unit 2023. An example of the environmental elements stored in the environmental element storage unit 2023 is shown in FIG.

[0025] In Figures 4 and 5, environmental elements A include the elements of "funder," "challenge method," "method of evaluating achievement level," "method of support," "incentive," "incentive allocation (in case of success)," "incentive allocation (in case of failure)," and "progress display in the app."

[0026] Moreover, the environmental element B is a medium item for each environmental element A, and is an element corresponding to each environmental element A. The environmental element C is a small item for each environmental element B, and is an element corresponding to each environmental element B.

[0027] The "funder" is a person who provides the funds for the incentive that is generated when the challenge is achieved. For example, the funder may be the person himself / herself, a family member, or a related person in the company.

[0028] The "challenge method" indicates how the user will tackle the challenge. For example, when the environmental element B is "self", the achievement is judged based on the result of one's own effort. When the environmental element B is "competition", the achievement is judged in conjunction with the results of other users. In FIG. 4, it is assumed that a company provides challenges to multiple employees as part of employee benefits, and the elements of "intra-department" and "departmental competition" are provided in the environmental element C. For example, when the challenge method is "competition" and "intra-department", a common challenge is carried out by the members of a certain department, and each member is ranked, and if the rank is above a certain level, the challenge is judged to be achieved. In "departmental competition", a department including at least one member can be made into a team and compete by team. When the environmental element B is "cooperation", the purpose is not for multiple people to compete, but for each member to carry out their own challenge and for everyone to achieve their challenge. In other words, if even one person fails as a joint responsibility, the challenge of all is considered to be a failure.

[0029] "Method of assessing achievement level" refers to a method of assessing achievement level of a challenge. For example, suppose there is a challenge to walk a specified number of steps per day for a specified number of days over a certain period of time. In that case, there are the following examples of methods of assessing achievement level. (A): A method of calculating achievement level by dividing the number of days achieved by the challenge period, i.e., (number of days achieved) / (specified number of days) x 100. (B): A method of calculating achievement level by referring to the step count records for each day starting with the record with the greatest number of steps for a specified number of days, and dividing the total of those step counts by the total number of steps if the goal is achieved.

[0030] For example, suppose there is a challenge to "walk 8,000 steps or more per day for 20 days or more out of 30 days." Here, suppose that during the period, you walked 4,000 steps for 10 days, 6,000 steps for 10 days, and 10,000 steps for 10 days. In this case, according to method (A) above, the achievement level would be (10 days) / (20 days) x 100 = 50%. Furthermore, according to method (B) above, considering that the total number of steps is 8,000 x 20 = 160,000 steps, the achievement level would be (6,000 steps x 10 days + 8,000 steps x 10 days) / 160,000 steps = 87.5%.

[0031] Also, if the challenge is different, other evaluation methods can be used. For example, suppose there is a challenge to improve one's 100-meter sprint time. In this case, the evaluation method is to evaluate the achievement level based on the rate of time reduction. Specifically, the achievement level is calculated as (number of seconds reduced) / (initial number of seconds) x 100. For example, if a person with a 100-meter sprint time of 15.0 seconds improves it to 14.5 seconds, the time reduction rate (achievement level) = 0.5 seconds / 15.0 seconds = 3.3%.

[0032] Such an "achievement evaluation method" is stored as environmental element C. Here, two environmental elements, method A and method B, corresponding to the above methods (A) and (B), are shown, but the evaluation methods can be added or changed appropriately depending on the challenge. Note that the "achievement evaluation method" does not need to be changed for each user, but may be associated in advance with each challenge in the challenge information described later in FIG. 7 so that a common achievement evaluation method is used for each user.

[0033] The "way of cheering" is used to set the "supporter", "support method" and "support frequency" who will support the user's challenge. Various options as shown in FIG. 4 are provided as environmental elements C.

[0034] In the "incentive" of FIG. 5, the "type", "amount" and "decrease in value over time" of the incentive that can be acquired when the user completes the challenge are set. In the "type", the form of providing the incentive is set as an environmental element C. In the "amount", the amount provided as the incentive or the amount of the goods or the like provided is set as an environmental element C. In the "decrease in value over time", the incentive is set to decrease when the completion of the challenge is delayed. For example, the environmental element C includes the elements "do not decrease", "decrease (20% / week)" and "decrease (50% / week)". Here, "decrease (20% / week)" means that, for example, when there is a challenge that is performed over four weeks, the amount of the incentive is decreased by 20% of the initial amount every week. By decreasing the incentive over time, the user can be motivated to actively perform the challenge.

[0035] "Incentive allocation (success)" and "Incentive allocation (failure)" set the incentive allocation method. Various options as shown in Figure 5 are provided as environmental elements C.

[0036] "Displaying progress in the app" sets the display mode of the progress of each user's challenge in the app installed on the user's device. An environmental element is provided for setting the display mode for competitors and supporters. Some people feel motivated by showing the progress of their challenge to competitors and supporters.

[0037] "Products and services used" are products, etc. used to complete a challenge. Here, as an example, environmental element C "supplements" and "training menu" included in environmental element B "exercise support" are shown. For example, in a challenge aimed at losing weight, the protein that is recommended for the user to use during the challenge period is included as an environmental condition as a supplement. Or, a menu of strength training for 10 minutes once a day is included as an environmental condition.

[0038] When a mail-order company recommends a product to a customer user, if the company can recommend the optimal product that the user desires, it will lead to a purchase. Or, even if the user does not feel that the product is valuable when it is recommended, if continuing to use the product has an effect such as weight loss, the reliability of the mail-order company's recommendation will increase, leading to a next purchase. It can be said that product recommendations that take into account not only the user's temporary evaluation of the product, but also the value the user will gain from the experience of using the product are optimal for the user.

[0039] For example, if user A with certain attributes takes protein A and does 10 minutes of strength training once a day, and the achievement rate of a certain goal (e.g., losing 2 kg of weight) is 80% (weight loss of 1.6 kg), then if a mail-order company recommends to user B, who has the same attributes as user A, that he or she take the same protein A and do 10 minutes of strength training once a day, user B may be able to lose weight, and this recommendation can be said to be highly satisfying to user B. In addition, when making a recommendation to user B, it may also be possible to tell that the achievement rate of the goal of losing 2 kg is estimated to be 80%.

[0040] If User B's actual weight loss was 1.8 kg, the achievement rate would be 90%, and the learning model can be trained using learning data such as User B's attributes, challenge environmental conditions (taking supplement A, strength training once a day for 10 minutes), and achievement rate of 90%. In this way, recommendations calculated based on evidence learned from challenges actually undertaken by many other users can help users to arrive at the value they are looking for without being overly misled by product designs and advertisements, ultimately resulting in high customer satisfaction.

[0041] An environmental condition for a certain challenge is created by combining each environmental element C. Also, the environmental condition does not need to include environmental elements C corresponding to all environmental elements A, and may be information that includes only some of the environmental elements.

[0042] Moreover, each item of the environmental element C is merely an example, and is not limited to those shown here. For example, for the distribution of "incentive" shown in FIG. 5, items of the environmental element may be provided in smaller monetary units. Furthermore, the "incentive distribution" is not limited to the distribution ratio exemplified here. Furthermore, it may be possible to set the incentive distribution to be performed among more parties involved, rather than distribution between two persons. Furthermore, for the "progress display in the app", items other than those shown here may be provided for the target for displaying the progress in the environmental element B and the display method in the environmental element C.

[0043] The suitability storage unit 2024 stores the suitability acquired by the suitability acquisition unit 207 described below, and attribute information and environmental condition candidates associated with the suitability, as learning data.

[0044] The attribute information reference unit 203 acquires user attribute information from the user terminal 102. Alternatively, the attribute information reference unit 203 accesses an information processing device (not shown) that stores and manages user attribute information via the network N, and acquires the attribute information of the target user.

[0045] The environmental condition generating unit 204 generates a challenge environmental condition based on the attribute information as an environmental condition that is a set of multiple environmental elements related to the environment in which the challenge is performed and stored in the environmental element storage unit 2023. The environmental condition generating unit 204 receives the attribute information as an input, and generates the challenge environmental condition based on a learning model M2 that is trained to output environmental conditions and stored in the learning model storage unit 2021.

[0046] The environment candidate generating unit 205 generates an environmental condition candidate including a plurality of environmental conditions by referring to a plurality of environmental elements. The environment candidate generating unit 205 is a part that functions during the generation process of the learning model M2.

[0047] The environmental element presentation unit 206 refers to a plurality of environmental elements and presents the plurality of environmental elements to the user. The environmental element presentation unit 206 generates information used to generate a screen for selecting each environmental element on the display unit 1022 of the user terminal 102, and transmits the information to the user terminal 102.

[0048] The suitability acquisition unit 207 presents multiple environmental condition candidates to a user involved in the generation of the learning model M2, and acquires the suitability of the challenge of the user under the multiple environmental condition candidates to the user. The suitability is the degree of achievement when a user who is presented with a certain challenge environmental condition actually performs a challenge under the challenge environmental condition, or the acceptability indicating the degree to which the user can accept the challenge environmental condition and perform the challenge with satisfaction.

[0049] The achievement level reference unit 208 accesses the challenge management device 103 via the communication unit 201 to refer to the achievement level managed by the challenge management device 103. There are achievement levels calculated from the results of a user's actual challenge (actual achievement level) and estimated achievement levels (estimated achievement level) described below. The achievement level referred to by the achievement level reference unit 208 is the actual achievement level.

[0050] The learning data storage unit 209 stores a set of the achievement level, the challenge environmental condition, and the attribute information as learning data in the memory unit 202. The learning data stored by the learning data storage unit 209 is used to generate a learning model M1.

[0051] The learning model generation unit 210 generates a learning model M1 based on the learning data, which is trained to receive attribute information as input and to output the challenge environmental conditions and the estimated achievement level.

[0052] In addition, the learning model generation unit 210 performs machine learning using the environmental condition candidates, the degree of conformance, and the attribute information as learning data, and generates a learning model M2 that is trained to use the attribute information as input and the challenge environmental condition as output. The generation process of the learning models M1 and M2 will be described later. When the degree of conformance is the degree of achievement, M1 and M2 are the same.

[0053] 6 shows a block diagram of the challenge management device 103. The challenge management device 103 has a communication unit 601, a storage unit 602, an attribute information reference unit 603, a challenge environmental condition acquisition unit 604, a judgment information acquisition unit 605, and an achievement level calculation unit 606. Each unit of the challenge management device 103 can be realized, for example, by using a storage area such as a memory in the challenge management device 103, or by a processor executing a program stored in the storage area.

[0054] The communication unit 601 controls communication between the challenge management device 103 and external terminals and information processing devices via the network N.

[0055] The storage unit 602 stores various information used in the processing in the challenge management device 103. The storage unit 602 includes a challenge storage unit 6021 and a model storage unit 6022.

[0056] The challenge storage unit 6021 displays challenge information as shown in Fig. 7. The user performs these challenges under environmental conditions set by the challenge management device 103. Note that the challenge information may be associated in advance with a method of evaluating the achievement level of each challenge. Furthermore, the model storage unit 6022 stores the learning model M1 generated by the learning model generation device 101 after being acquired from the learning model generation device 101.

[0057] The attribute information reference unit 603 acquires the attribute information of the user. The attribute information reference unit 603 acquires the attribute information of the user from the user terminal 102. Alternatively, the attribute information reference unit 603 accesses an information processing device (not shown) that stores and manages the attribute information of the user via the network N, and acquires the attribute information of the target user.

[0058] The challenge environmental condition acquisition unit 604 inputs the user's attribute information into the learning model M1 and acquires the user's challenge environmental conditions.

[0059] The judgment information acquisition unit 605 acquires judgment information used to judge achievement, including calculation of the achievement level of a challenge of a user who executes a challenge for which challenge environmental conditions are set. For example, in the case of a challenge related to the number of steps, the number of steps is acquired from the user terminal 102. For example, if the user terminal 102 is a smartphone equipped with a step count measurement function, that information can be used to measure the number of steps. In addition, the judgment information may be acquired from a terminal other than the user terminal 102. For example, when a user other than the user who executes the challenge makes the judgment, the judgment information can be acquired from a terminal of the judge, the judgment organization, or the like.

[0060] The achievement level calculation unit 606 calculates the user's achievement level of the challenge based on the achievement level evaluation method included in the challenge environmental conditions.

[0061] The generation process of the learning model M2 in the learning model generation device 101 will be described with reference to FIG. 8 and FIG. 9 to FIG.

[0062] In step S 801 , the attribute information reference unit 203 acquires user attribute information based on the attribute elements stored in the attribute element storage unit 2022 .

[0063] In step S 802 , the environmental condition generating unit 204 acquires environmental elements from the environmental element storage unit 2023 .

[0064] In step S803, the matching degree acquiring unit 207 judges whether or not to acquire the matching degree as the achievement degree. This judgment is made based on information preset by the administrator of the learning model generating device 101.

[0065] If an affirmative decision is made in step S803, in step S804, the candidate environment generating unit 205 generates a plurality of candidate environmental conditions by randomly generating environmental conditions based on environmental elements.

[0066] In step S805, the compatibility acquisition unit 207 transmits the generated environmental condition candidates to the user terminal 102 and the challenge management device 103.

[0067] In step S806, the degree of conformance acquisition unit 207 acquires, for each of the environmental condition candidates, the degree of achievement calculated by the challenge management device 103 as a result of the user performing the challenge for each of the environmental condition candidates.

[0068] In step S807, the suitability acquisition unit 207 stores a set of the degree of achievement as the degree of achievement, the environmental condition candidate, and the attribute information as learning data in the suitability storage unit 2024. A predetermined threshold may be set for the degree of achievement, and a storage target for the learning data may be determined based on the threshold. For example, a minimum degree of achievement may be set as a predetermined threshold for the degree of achievement, and if the degree of achievement exceeds the minimum degree of achievement, it may be stored as learning data. Alternatively, from the viewpoint of preventing injuries in the 100m sprint, a challenge of "getting injured in the 100m sprint" may be made, and environmental conditions for practice in which the degree of achievement of the challenge of "getting injured in the 100m sprint" is 5% or less may be output. In this case, a maximum degree of achievement may be set as a predetermined threshold for the degree of achievement, and if the degree of achievement does not exceed the maximum degree of achievement, it may be stored as learning data.

[0069] If a negative decision is made in step S803, the suitability acquisition unit 207 decides in step S808 whether or not to acquire the suitability as the user's acceptance of the environmental condition.

[0070] If an affirmative decision is made in step S808, in step S809, the candidate environment generating unit 205 generates candidate environmental conditions by randomly generating environmental conditions based on environmental elements.

[0071] In step S810, the suitability obtaining unit 207 transmits the generated environmental condition candidates to the user terminal 102. Here, the reason why it is not necessary to transmit the environmental condition candidates to the challenge management device 103 is that the acceptance level is based on the subjective judgment of the user, and it is not necessary to use the achievement level calculated by the challenge management device 103.

[0072] In step S811, the suitability acquisition unit 207 acquires the acceptability of each environmental condition candidate from the user terminal 102. In acquiring the acceptability, the user terminal 102 may display the environmental conditions on the display unit 1022, and the user may input the acceptability for the environmental condition.

[0073] In step S812, the compatibility acquisition unit 207 stores a set of the acceptability as the compatibility, the environmental condition candidate, and the attribute information in the compatibility storage unit 2024 as learning data.

[0074] If a negative decision is made in step S808, the environmental element presenting unit 206 refers to the environmental element storage unit 2023 and transmits a list of environmental elements to the user terminal 102 in step S813.

[0075] In step S814, the suitability acquisition unit 207 acquires, from the user terminal 102, environmental conditions that are created by the user selecting environmental elements displayed on the user terminal 102.

[0076] In step S815, the compatibility acquisition unit 207 stores a pair of the attribute information and the acquired environmental condition in the compatibility storage unit 2024 as learning data.

[0077] In step S816, the learning model generation unit 210 generates the learning model M2 based on the stored learning data. A general machine learning method can be used to generate the learning model M2. For example, the learning model M2 is generated by learning a neural network that inputs attribute information and outputs environmental conditions and achievement or acceptance based on the learning data. The learning model M1 described later can also be generated using a method similar to that for generating the learning model M2, in which attribute information is input and achievement is output. The generated learning model is stored in the learning model storage unit 2021. The learning model M2 is configured to output challenge environmental conditions for each challenge content based on attribute information of the user who will perform the challenge.

[0078] Fig. 9 is an example of information indicating candidate environmental conditions and the degree of achievement for a certain challenge when the degree of conformance is the degree of achievement. The candidate environmental conditions shown are environmental condition A, environmental condition B, and environmental condition C. Note that for each environmental condition, a part of the environmental elements is shown as an example. In Fig. 9, the actual degree of achievement of the challenge under each environmental condition is recorded.

[0079] The environmental elements may include a training menu or the like required to achieve the challenge depending on the challenge content. For example, if the challenge content is to shorten the time for the 100-meter sprint, environmental conditions may be generated that include necessary muscle training items as environmental elements.

[0080] Fig. 10 is an example of information showing each environmental condition and the acceptance level when the suitability is the acceptance level. Fig. 11 is an example of information when the environmental condition is acquired based on the user's selection. In the case of Fig. 11, the acceptance level is set to 100% for convenience because it is based on the user's selection.

[0081] The process up to generation of the learning model M1 in the challenge support system 10 will be described with reference to Fig. 12. Here, it is assumed that the challenge is performed by a user operating the user terminal 102a.

[0082] In step S1201, the learning model generating device 101 performs the above-mentioned processing for generating the learning model M2.

[0083] In step S1202, the attribute information reference unit 603 acquires attribute information and information related to the challenge that the user will attempt from the user terminal 102a. The information related to the challenge is information for identifying the content of the challenge that the user will attempt.

[0084] The challenge management information stored in the challenge storage unit 6021 of the challenge management device is updated based on the information acquired by the attribute information reference unit 603. An example of the challenge management information is shown in FIG.

[0085] The challenge management information has the following items: "user name," "user ID," "attribute information," "challenge ID," "challenge name," "challenge environmental conditions," and "achievement level."

[0086] The "user name" is the user's name or nickname, etc. The "user ID" is an ID for identifying the user. The "attribute information" is the attribute information of each user. The "challenge ID" is an ID for identifying each challenge. The "challenge name" is the challenge name of the challenge information shown in FIG. 7. The content of the challenge that each user will attempt is identified based on the challenge name. The "challenge environmental conditions" are environmental conditions generated based on the attribute information of each user. Here, if challenges with "challenge IDs" of "CID001" to "CID004" are generated by the processing up to FIG. 12, each environmental condition is a condition generated using the learning model M2. The "achievement level" is the achievement level calculated based on the judgment information when each challenge is executed.

[0087] Based on the challenge-related information and attribute information acquired by the attribute information reference unit 603 in step S1202, the contents of "user name", "user ID", "attribute information", "challenge name", and "challenge ID" are recorded.

[0088] In step S1203, the challenge environmental condition acquisition unit 604 transmits information and attribute information related to the challenge to the learning model generation device 101. In step S1204, the environmental condition generation unit 204 generates challenge environmental conditions using the learning model M2 based on the acquired information and attribute information related to the challenge. Here, in the learning model M2, the challenge environmental conditions are generated by inputting the user's attribute information and outputting the environmental conditions with the highest degree of suitability.

[0089] In step S1205, the challenge environmental condition acquisition unit 604 acquires the generated challenge environmental condition. The challenge environmental condition acquisition unit 604 records the content of the "challenge environmental condition" of the challenge management information based on the acquired challenge environmental condition.

[0090] In step S1206, the challenge environmental condition acquisition unit 604 transmits the challenge environmental conditions to the user terminal 102a. In step S1207, determination information when the user executes the challenge under the challenge environmental conditions is acquired by the user terminal 102a.

[0091] In step S1208, the determination information acquisition unit 605 acquires the determination information from the user terminal 102a. In step S1209, the achievement level calculation unit 606 calculates the achievement level based on the determination information. The calculated achievement level is stored in the challenge management information in association with the target challenge.

[0092] The judgment information acquisition unit 605 may acquire judgment information from a terminal other than the user terminal 102a. In this case, the challenge environment condition acquisition unit 604 transmits the challenge environment conditions to the user terminal 102a, and the conditions are displayed on the user terminal 102a, but the acquisition of information necessary for judgment may not be performed. This also applies to the update process of the learning model M1 described later.

[0093] In step S 1210 , the achievement level reference unit 208 acquires the achievement level from the challenge management device 103 .

[0094] In step S1211, the learning data storage unit 209 stores a set of the achievement level, the challenge environmental condition, and the attribute information as learning data in the memory unit 202. A predetermined threshold may be set for the achievement level, and the learning data to be stored may be determined based on the threshold. For example, a minimum achievement level or a maximum achievement level may be set as a predetermined threshold for the achievement level, and if the achievement level exceeds the minimum achievement level or does not exceed the maximum achievement level, it may be determined that the achievement level is to be stored as learning data.

[0095] The processes up to this point can be performed for a plurality of users other than the user who operates the user terminal 102a. In step S1211, the learning model generation unit 210 generates a learning model M1 using the learning data collected from each user.

[0096] In step S1213, the learning model generation unit 210 transmits information about the learning model M1 to the challenge management device 103. In step S1214, the challenge management device 103 acquires information about the learning model M1 through the communication unit 601 and stores the learning model M1 in the model storage unit 6022.

[0097] Note that, although the generation of the learning model M1 is described here as an example in which the learning model M2 is used, the learning model M1 may also be generated, for example, based on challenge environmental conditions obtained from the user without using the learning model M2.

[0098] Furthermore, the learning data stored in the memory unit 202 by the learning data storage unit 209 may include external information such as an average achievement level by generation used to evaluate the achievement level.

[0099] 14 is a diagram for explaining the process of relearning the learning model M1 when the learning model M1 is generated by the learning model generation device 101 using the learning model M2. In this case, the learning model M1 is stored in the model storage unit 6022. Here, an update of the learning model M1 by a user operating the user terminal 102b performing a challenge will be explained.

[0100] In step S1401, the attribute information reference unit 603 acquires attribute information and information related to a challenge that the user will attempt from the user terminal 102b.

[0101] Based on the information about the challenge and the attribute information acquired by the attribute information reference unit 603 in step S1401, the contents of "user name", "user ID", "attribute information", "challenge name", and "challenge ID" are recorded. By the attribute information reference unit 603 recording each content, a new challenge is added to the challenge management information. An example of the challenge management information after the challenge has been added is shown in Fig. 15.

[0102] In step S1402, the challenge environmental condition acquisition unit 604 generates challenge environmental conditions based on the user's attribute information using the learning model M1 stored in the model storage unit 6022. Here, in the learning model M1, the user's attribute information is input, and the challenge environmental conditions are generated such that the environmental condition with the highest achievement level and the estimated achievement level are output. The challenge environmental condition acquisition unit 604 records the contents of the "challenge environmental conditions" in the challenge management information based on the generated challenge environmental conditions.

[0103] In step S1403, the challenge environmental condition acquisition unit 604 transmits the challenge environmental conditions to the user terminal 102b. In step S1404, determination information when the user executes a challenge under the challenge environmental conditions is acquired by the user terminal 102b.

[0104] In step S1405, the determination information acquisition unit 605 acquires the determination information from the user terminal 102b. In step S1406, the achievement level calculation unit 606 calculates the achievement level based on the determination information. The calculated achievement level is stored in the challenge management information in association with the target challenge.

[0105] In step S1407, the achievement level reference unit 208 acquires information about the challenge, attribute information, challenge environmental conditions, and the achievement level from the challenge management device 103.

[0106] In step S1408, the learning data storage unit 209 stores the set of the achievement level, the challenge environment condition, and the attribute information as new learning data for the learning model M1. A predetermined threshold may be set for the achievement level, and the learning data to be stored may be determined based on the threshold. For example, a minimum achievement level or a maximum achievement level may be set as a predetermined threshold for the achievement level, and if the achievement level exceeds the minimum achievement level or does not exceed the maximum achievement level, it may be determined that the achievement level is to be stored as learning data.

[0107] The processes up to this point can be performed for a plurality of users other than the user who operates the user terminal 102b. In step S1409, the learning model M1 can be updated using new learning data collected from each user.

[0108] In step S1410, the learning model generation unit 210 transmits information about the learning model M1 to the challenge management device 103. In step S1411, the challenge management device 103 acquires information about the learning model M1 through the communication unit 601 and stores the learning model M1 in the model storage unit 6022.

[0109] Based on the attribute information of the user, one challenge environmental condition may be generated, or multiple challenge environmental conditions may be generated. A case where multiple challenge environmental conditions are generated will be described. For example, when the learning model M1 generates a certain challenge environmental condition A0, the challenge environmental condition acquisition unit 604 generates challenge environmental conditions A1 and A2 by changing a part or all of the challenge environmental condition A0. Changing a part or all of the challenge environmental condition A0 includes randomly changing the environmental elements A, B, and C, or randomly changing the environmental elements B and C for a specific environmental element A.

[0110] That is, the challenge environmental condition acquisition unit 604 generates challenge environmental conditions A1 and A2 by changing a part or all of the challenge environmental conditions A0 generated by the learning model M1 based on the challenge environmental conditions A0, and transmits the challenge environmental conditions A0 and the challenge environmental conditions A1 and A2 to the user. The achievement level reference unit 208 acquires the achievement level when a challenge is actually performed under the challenge environmental conditions A0 and the challenge environmental conditions A1 and A2. The learning data storage unit 209 stores a set of the achievement level, the challenge environmental conditions, and the attribute information for each challenge environmental condition as new learning data for the learning model M1. A predetermined threshold value may be set for the achievement level, and the learning data storage target may be determined based on the threshold value. For example, a minimum achievement level or a maximum achievement level may be set as a predetermined threshold value for the achievement level, and if the achievement level exceeds the minimum achievement level or does not exceed the maximum achievement level, it may be stored as learning data.

[0111] When a user challenges the generated challenge environmental conditions A0, A1, and A2, and a challenge environmental condition with a higher achievement level than the challenge environmental condition A0 is generated, the challenge environmental condition becomes the learning data of the learning model M1. Therefore, the environmental conditions generated by the learning model M1 can be made to further increase the achievement level of the user.

[0112] 13 and 15, the setting of challenge environment conditions using the learning model M2 and the setting of challenge environment conditions using the learning model M1 in the challenge support system 10 will be described.

[0113] In FIG. 13, user A takes on challenge 1 in an environment with challenge environmental conditions "environmental condition A" generated by learning model M2 based on attribute information A. The achievement rate of user A in the challenge is 70%. User B takes on the same challenge as user A in an environment with challenge environmental conditions "environmental condition A" generated by learning model M2 based on attribute information B. The achievement rate of user B in the challenge is 90%.

[0114] In this case, it is assumed that the achievement level of user A is not sufficient. In this case, it is not preferable to recommend challenge environment conditions based on learning model M2 to user A or a user having attribute information similar to that of user A in terms of improving the achievement level.

[0115] Therefore, assume that the process in FIG. 12 has yielded a learning model M1 that generates environmental conditions suitable for a user with attribute information similar to that of user A. When learning model M1 is used, the challenge environmental conditions are set as "environmental conditions B" based on the same attribute information for a challenge with user A's challenge ID "CID101." When the challenge environmental conditions are set in this manner, the achievement level is 90%. In other words, learning model M1, which generates and updates a model based on the actual achievement level, can provide more appropriate challenge environmental conditions to the user.

[0116] In addition, by using the learning model M1, it is possible to consider appropriate values ​​for variable attribute information among the attribute information of the user. The process in this case will be described with reference to Figs. 16 to 20.

[0117] 16 is a diagram showing an outline of this process. In step S1601, the attribute information reference unit 603 acquires, from the user terminal 102, information related to a challenge that the user will attempt and attribute information.

[0118] In step S1602, the challenge environmental condition acquisition unit 604 generates challenge environmental conditions based on the learning model M1.

[0119] In step S1603, the challenge environmental condition acquisition unit 604 acquires an estimated achievement level (estimated achievement level) of the generated challenge environmental conditions.

[0120] In step S1604, the challenge environmental condition acquisition unit 604 transmits the challenge environmental conditions and the estimated achievement level to the user terminal 102. In step S1605, the display unit 1022 displays the estimated achievement level for certain attribute information.

[0121] For example, suppose there is attribute information for the work environment attribute item A with a welfare budget of "5,000 yen" as shown in Fig. 17. Note that appropriate information may also be input as values ​​for other items.

[0122] In this case, the challenge environmental conditions and estimated achievement level generated using the learning model M1 are as shown in FIG.

[0123] Next, the value of the attribute information is changed to “4000 yen.” In this case, the challenge environmental conditions and estimated achievement level generated using the learning model M1 are as shown in FIG.

[0124] The estimated achievement levels in Fig. 18 and Fig. 20 are 80% in Fig. 18 and 78% in Fig. 20, respectively. In this case, it is considered that the reduction in the estimated achievement level is small compared to the reduction in the welfare program budget. Therefore, for example, when the attribute information of employees can be managed by a company, a person in charge at the company can operate the user terminal 102 and use the attribute information of the employees to consider values ​​of the attribute information that ensure a certain degree of achievement while suppressing the welfare program budget. In addition, the above consideration may be made possible by a mode in which an average value of the estimated achievement levels of each employee is displayed when the welfare program budget of each attribute information of multiple employees is changed collectively.

[0125] The present embodiment has been described above. The learning model generation device 101 according to the present embodiment includes an attribute information reference unit 203 that references attribute information indicating the attributes of a user who has a challenge to be completed, an environmental condition generation unit 204 that generates a challenge environmental condition as an environmental condition that is a set of a plurality of environmental elements stored in a storage unit regarding an environment in which the challenge is performed based on the attribute information, an achievement level reference unit 208 that references the achievement level of the challenge when the challenge is performed by the user under the challenge environmental condition, a learning data storage unit 209 that stores a set of the achievement level, the challenge environmental condition, and the attribute information in the storage unit 202 as learning data, and a learning model generation unit 210 that generates a learning model M1 that is trained to use the attribute information as input and the challenge environmental condition as output based on the learning data.

[0126] The environmental condition generating unit 204 generates a certain challenge environmental condition based on the attribute information referred to by the attribute information referring unit 203. The achievement level referring unit 208 refers to the achievement level of the user when the user executes the challenge under the challenge environmental condition. The learning model generating unit 210 generates a learning model M1 using learning data including the evaluation result of the actual challenge, that is, the achievement level. For example, the learning model M1 functions as a learning model capable of presenting a challenge environment in which the user is likely to achieve the challenge by outputting the challenge environmental condition that maximizes the achievement level.

[0127] Furthermore, the learning model generation unit 210 may generate a learning model that is trained based on the learning data to input attribute information and output an estimated achievement level of the challenge. This makes it possible to display the estimated achievement level to the user. The user who is displayed the estimated achievement level can know the likelihood of achieving the challenge using the numerical value as a guide, which may enable the user to actively work on the challenge. Thus, the learning model M1 functions as a learning model that can present a challenge environment in which the user is likely to achieve the challenge.

[0128] Furthermore, in the learning model generation device 101, the environmental condition generation unit 204 receives attribute information as input, and outputs challenge environmental conditions based on the learning model M2 that has been trained to output environmental conditions and stored in the storage unit 202. This allows the learning model M2 to be trained to present challenge environmental conditions, thereby effectively learning the learning model M1 based on the achievement level.

[0129] In addition, in the learning model generation device 101, the attribute information reference unit 203 further includes an environmental candidate generation unit 205 that references second attribute information of a second user different from the first user and references multiple environmental elements to generate environmental condition candidates including multiple environmental conditions, and a suitability acquisition unit 207 that presents the multiple environmental condition candidates to the second user and acquires the suitability of the multiple environmental condition candidates for the second user, and the environmental condition generation unit 204 generates challenge environmental conditions based on a learning model M2 that is machine-learned to use the environmental condition candidates, the suitability, and the second attribute information as second learning data and relates the second attribute information to the environmental condition candidates.

[0130] As a result, the learning model M2 is machine-trained using learning data including the suitability of the challenge of the second user to the second user in a plurality of candidate environmental conditions presented to the second user. The learning model M2 is a model that presents challenge environmental conditions to the user in order to generate the learning model M1. By having the learning model M2 learn in advance to present challenge environmental conditions that suit the user, it is possible to effectively learn the learning model M1 based on the achievement level.

[0131] In addition, in the learning model generation device 101, the compatibility acquisition unit 207 acquires, as compatibility, the achievement levels of multiple challenges when the second user performs multiple challenges in each of multiple environmental condition candidates. This allows the learning model M2 to learn using the achievement levels when the user actually performs the challenges, so that the learning model M2 can output the challenge environmental conditions more appropriately.

[0132] In addition, in the learning model generation device 101, the learning model generation unit 210 further re-learns the learning model M2 based on the teacher data. This enables the learning model M2 to present more appropriate challenge environmental conditions according to the user's attribute information, so that the final presentation of challenge environmental conditions using the learning model M1 becomes more appropriate.

[0133] In addition, in the learning model generation device 101, the environmental candidate generation unit 205 generates presentation information for presenting multiple environmental condition candidates to a second user, and the suitability acquisition unit 207 acquires the user's acceptance of the multiple environmental condition candidates based on the presentation information as suitability.

[0134] This allows the challenge environment conditions that are acceptable to the user to be determined based on the acceptability without the user having to actually perform the challenge. When the acceptability is high, the user is more likely to take action to complete the challenge, and the likelihood of completing the challenge increases. By using the acceptability, the learning model M2 is trained without the user having to actually perform the challenge, and the learning model M1 based on the challenge environment conditions can be effectively trained.

[0135] In addition, the learning model generation device 101 further includes an environmental element presentation unit 206 that presents a plurality of environmental elements to the second user, and the environmental condition generation unit 204 generates challenge environmental conditions based on a learning model M2 that has been machine-learned using the environmental conditions, which are a set of environmental elements selected by the second user, and the second attribute information as third teacher data.

[0136] This makes it possible to obtain challenge environmental conditions that are acceptable to the user without the user having to actually perform the challenge. In the case of environmental conditions selected by the user himself, the user is more likely to take action to complete the challenge, and the likelihood of completing the challenge is increased. By training the learning model M2 without requiring the user to actually perform the challenge, it is possible to effectively train the learning model M1 based on the challenge environmental conditions.

[0137] The challenge support system 10 also includes a learning model generation device 101 and a challenge management device 103 that manages challenges.

[0138] The challenge management device 103 includes a model storage unit 6022 in which the learning model M1 generated by the learning model generation device 101 is stored, an attribute information reference unit 603 that references the attribute information of the user, a challenge environmental condition acquisition unit 604 that inputs the attribute information to the learning model M1 and acquires the challenge environmental conditions of the user, a judgment information acquisition unit 605 that acquires judgment information for making a judgment on the achievement conditions of the challenge, and an achievement level calculation unit 606 that calculates the achievement level of the challenge under the challenge environmental conditions based on the judgment information. In the learning model generation device 101, the achievement level reference unit 208 acquires the achievement level under the challenge environmental conditions by referring to the achievement level calculated by the achievement level calculation unit 606.

[0139] The challenge support system 10 enables the generation of a learning model by the learning model generation device 101, and the generation of challenge environment conditions using the learning model and management of the challenge by the challenge management device 103. This allows a challenge environment in which the user is likely to complete the challenge to be presented, and the user can be encouraged to perform the challenge, thereby encouraging the completion of the challenge. The functions of the learning model generation device 101 and the challenge management device 103 may be integrated into one device.

[0140] The challenge environment condition acquisition unit 604 may also acquire an estimated achievement level in the challenge environment conditions from the learning model M1 capable of outputting an estimated achievement level of the challenge based on the attribute information, and generate information for displaying the estimated achievement level so that the user can view it. This makes it possible to display the estimated achievement level to the user. This enables the user, who uses the estimated achievement level as a guide, to actively work on the challenge, and encourages the achievement of the challenge.

[0141] In this embodiment, the program for operating the learning model generating device 101 can be stored in a storage medium. The storage medium storing the program may be a non-transitory storage medium (Non-transitory computer readable medium). The non-transitory storage medium is not particularly limited, and may be, for example, a storage medium such as a USB memory or a CD-ROM.

[0142] The above-described embodiment is intended to facilitate understanding of the present invention, and is not intended to limit the present invention. The elements and conditions of the embodiment are not limited to those exemplified, and may be changed as appropriate. [Explanation of symbols]

[0143] 10... Challenge support system, 101... Learning model generation device, 102a, 102b... User terminal, 103... Challenge management device, 203... Attribute information reference unit, 204... Environmental condition generation unit, 205... Environment candidate generation unit, 206... Environmental element presentation unit, 207... Conformity acquisition unit, 208... Achievement level reference unit, 209... Learning data storage unit, 210... Learning model generation unit

Claims

1. An attribute information reference unit that references attribute information indicating attributes of a user; an environmental condition generation unit that generates a challenge environmental condition, which is a set of a plurality of environmental elements, the plurality of environmental elements including information that sets an incentive for a user's challenge and an allocation of a reward based on the incentive; an achievement level reference unit that references an achievement level of the challenge when the challenge is executed by the user under the challenge environment conditions; a learning data storage unit that stores a set of the achievement level, the challenge environmental condition, and the attribute information in a storage unit as learning data in association with the content of the challenge; A learning model generation device comprising: a learning model generation unit that generates a learning model based on the learning data, the learning model being trained to use the attribute information as input and the challenge environmental conditions as output.

2. The learning model generation device according to claim 1, The learning model generation unit is a learning model generation device that generates a learning model based on the learning data, which is trained to use the attribute information as input and to output the estimated degree of achievement of the challenge.

3. The learning model generation device according to claim 1 or 2, The user is a first user, The attribute information is first attribute information of the first user, The learning model is a first learning model, The learning data is first learning data, The attribute information reference unit further refers to second attribute information of a second user, The learning model generation device includes: an environment candidate generation unit that generates a plurality of environmental condition candidates that are candidates for the challenge environmental condition by randomly combining the plurality of environmental elements with reference to the plurality of environmental elements; a suitability acquisition unit that presents the plurality of environmental condition candidates to the second user and acquires a suitability of the challenge of the second user to the plurality of environmental condition candidates, the environmental condition generation unit inputs the second attribute information into a second learning model to generate the challenge environmental conditions; The second learning model is machine-trained using the environmental condition candidate, the degree of compatibility, and the second attribute information as second learning data, and outputs a challenge environmental condition included in the multiple environmental condition candidate in response to input of attribute information.

4. The learning model generating device according to claim 3, A learning model generation device, wherein the compatibility acquisition unit acquires as the compatibility the degree of completion of the multiple challenges when the multiple challenges in each of the multiple candidate environmental conditions are performed by the second user.

5. The learning model generating device according to claim 4, A learning model generation device, wherein the learning model generation unit further re-learns the second learning model based on the learning data.

6. The learning model generating device according to claim 3, the environment candidate generation unit generates presentation information for presenting the plurality of environmental condition candidates to the second user; The compatibility acquisition unit acquires the second user's acceptability for the multiple environmental condition candidates based on the presentation information as the compatibility from the user terminal of the second user, in a learning model generation device.

7. The learning model generating device according to claim 2, The user is a first user, The attribute information is first attribute information of the first user, The learning data is first learning data, The attribute information reference unit refers to second attribute information of a second user, an environmental element presenting unit that presents the plurality of environmental elements to the second user, the environmental condition generation unit inputs the second attribute information into a second learning model to generate the challenge environmental conditions; The second learning model is machine-trained using the challenge environmental conditions, which are a set of environmental elements selected by the second user, and the second attribute information as third learning data, and outputs challenge environmental conditions in response to input of attribute information.

8. The learning model generation device according to any one of claims 1 to 7, A learning model generation device, wherein the attribute information reference unit references attribute information including information regarding the user's cognitive bias.

9. A learning model generating device according to any one of claims 1 to 8, a challenge management device for managing the challenge; Equipped with The attribute information reference unit of the learning model generation device is a first attribute information reference unit, The challenge environmental conditions include information on how to evaluate the challenge; The challenge management device a model storage unit in which the learning model generated by the learning model generation device is stored; a second attribute information reference unit that references the attribute information of the user; a challenge environment condition acquisition unit that inputs the attribute information into the learning model and acquires the challenge environment condition of the user; a determination information acquisition unit that acquires determination information for determining whether or not the challenge has been met; an achievement level calculation unit that calculates an achievement level of the challenge under the challenge environment conditions based on the determination information and the evaluation method, In the learning model generation device, A challenge support system, wherein the achievement level reference unit refers to the achievement level calculated by the achievement level calculation unit to obtain the achievement level under the challenge environmental conditions.

10. The challenge support system according to claim 9, In the challenge management device, A challenge support system in which the challenge environmental condition acquisition unit acquires the estimated degree of achievement under the challenge environmental conditions from the learning model capable of outputting an estimated degree of achievement of the challenge based on the attribute information, and generates information for displaying the estimated degree of achievement so that the user can view it.

11. On the computer, An attribute information reference process for referring to attribute information indicating attributes of a user; An environmental condition generation process for generating a challenge environmental condition, which is a set of a plurality of environmental elements, the plurality of environmental elements including information for setting an incentive for a user's challenge and an allocation of a reward based on the incentive; an achievement level reference process for referring to an achievement level of the challenge when the challenge is executed by the user under the challenge environment conditions; a learning data storage process for storing a set of the achievement level, the challenge environmental condition, and the attribute information in a storage unit as learning data in association with the content of the challenge; A program that executes a learning model generation process that generates a learning model based on the learning data, the learning model being trained to use the attribute information as input and the challenge environmental conditions as output.

Images