Meal proposal device, meal proposal method, and meal proposal program

Abstract

To provide a meal proposal device capable of proposing meals to a target user with clearly stated reasons for the selection, a meal proposal method, and a meal proposal program.SOLUTION: A meal proposal system comprises a terminal device used by users such as object employees who are the meal proposal destination, a meal proposal device, and a storage device. A meal proposal device 100 includes: an acquisition unit which acquires subject-related data related to a subject targeted for meal proposal; a meal proposal unit which determines a meal to be proposed to the subject using the subject-related data; a rationale information generation unit which generates rationale information indicative of the basis for the proposed meal; and a presentation unit which presents the proposed meal and the rationale information.SELECTED DRAWING: Figure 3

Description

[Technical field]

[0001] The present invention relates to a meal recommendation device, a meal recommendation method, and a meal recommendation program. [Background technology]

[0002] Patent Document 1 describes a centralized health and diet management system that notifies employees of recommended menus based on employee data. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2002-334158 A Summary of the Invention [Problem to be solved by the invention]

[0004] Patent Document 1 only notifies the employee of the recommended meal, and does not clearly explain the reason for the selection. Even if the user looks at the suggested meal, it is difficult for them to understand why the meal was recommended and what criteria or calculation logic it is based on. This problem can also occur when suggesting meals to non-employees.

[0005] One aspect of the present invention has been made in consideration of the above problems, and its purpose is to suggest a meal to a subject after clearly explaining the reason for the selection. [Means for solving the problem]

[0006] One aspect of the meal suggestion device has an acquisition unit that acquires subject-related data related to a subject to which meals are suggested, a meal suggestion unit that uses the subject-related data to determine a meal to suggest to the subject, a basis information generation unit that generates basis information indicating the basis for suggesting the meal, a presentation unit that presents the suggested meal and the basis information, and a meal suggestion unit.

[0007] One aspect of the meal suggestion method includes at least one processor obtaining subject-related data related to a subject to whom a meal is to be suggested, using the subject-related data to determine a meal to suggest to the subject, generating evidence information indicating a basis for suggesting the meal, presenting the suggested meal, and presenting the evidence information.

[0008] One aspect of the program records a program that causes a computer to execute the following processes: acquiring employee-related data; selecting a meal to suggest to the employee based on the employee-related data; calculating the basis for the selection; and outputting the meal and the basis. Effect of the Invention

[0009] According to one aspect of the present invention, meals can be suggested to employees after clearly explaining the reasons for their selection. [Brief description of the drawings]

[0010] [Figure 1] 1 shows the overall configuration of a meal recommendation system according to the present disclosure. [Diagram 2] 2 shows the hardware configuration of the meal recommendation device. [Diagram 3] FIG. 1 is a block diagram of a meal suggestion device according to the present disclosure. [Figure 4] Examples of employee-related data according to the present disclosure are shown below. [Diagram 5] 1 shows an example presentation according to the present disclosure. [Figure 6] 1 shows a flowchart according to the present disclosure. [Figure 7] FIG. 1 is a block diagram of a meal suggestion device according to the present disclosure. [Figure 8] 1 shows an example presentation according to the present disclosure. [Figure 9] 1 shows a flowchart according to the present disclosure. [Figure 10] 1 shows an example of dietary information according to the present disclosure. [Figure 11] 1 shows an example presentation according to the present disclosure. [Figure 12] FIG. 1 is a block diagram of a meal suggestion device according to the present disclosure. [Figure 13] FIG. 1 is a block diagram of a meal suggestion device according to the present disclosure. [Figure 14] 1 shows an example presentation according to the present disclosure. [Figure 15] 1 shows a flowchart according to the present disclosure. [Figure 16] FIG. 1 is a block diagram of a meal suggestion device according to the present disclosure. [Figure 17] 1 shows an example presentation according to the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, a preferred embodiment of the present disclosure will be described with reference to the drawings. In this embodiment, an example in which the target of meal suggestions is an employee will be described, but the target of meal suggestions in the present invention is not limited to employees. For example, the target may be a user who attends a fitness gym, an inpatient in a hospital, or a resident of an elderly care facility.

[0012] First Embodiment (Overall composition) In the first embodiment, a meal recommendation system that suggests meals to employees after clarifying the reasons for their selection will be described.

[0013] 1 shows the overall configuration of a meal recommendation system 1000 according to the first embodiment. The meal recommendation system includes a terminal device 20 used by a user such as a target employee 10 to whom meals are to be recommended, a meal recommendation device 100, and a storage device 5.

[0014] The meal suggestion device 100 generates suggested meal information 40 and basis information 50 for the target employee 10 based on the employee-related data 30, and transmits them to the terminal device 20. Note that the employee-related data 30 corresponds to target-person-related data for the target person.

[0015] (Hardware configuration) 2 is a block diagram showing a hardware configuration of the meal suggestion device 100. As shown in the figure, the meal suggestion device 100 includes a processor 1, an input / output interface 2, a ROM (Read Only Memory) 3, a RAM (Random Access Memory) 4, and a storage device 5. Each component is connected to each other via a bus 6, for example.

[0016] The processor 1 is a computer such as a CPU (Central Processing Unit), and executes a meal recommendation program prepared in advance to control the entire meal recommendation device 100. Specifically, the processor 1 may be a CPU, a GPU (Graphics Processing Unit), a DSP (Digital Signal Processor), an MPU (Micro Processing Unit), an FPU (Floating Point number Processing Unit), a PPU (Physics Processing Unit), a TPU (Tensor Processing Unit), a quantum processor, a microcontroller, or a combination of these.

[0017] The processor 1 also loads programs stored in the ROM 3, storage device 5, etc. The processor 1 then executes each process coded in the program. The processor 1 also functions as a part or all of the meal suggestion device 100. The processor 1 may then execute processes or instructions in a flowchart, which will be described later, based on the program.

[0018] The input / output interface 2 is an interface through which the meal suggestion device 100 transmits and receives data to and from other devices. For example, the meal suggestion device 100 acquires various data on the target employee 10 by communicating with the terminal device 20 via the input / output interface. In addition, the meal suggestion device 100 transmits the generated suggested meal information 40 and grounds information 50 to the terminal device 20 via the input / output interface 2.

[0019] The ROM 3 stores various programs executed by the processor 1. In addition, the RAM 4 is used as a working memory while the processor 1 is executing various processes.

[0020] The storage device 5 is a non-volatile, non-transient storage device. For example, the storage device 5 may be a disk-shaped recording medium, a semiconductor memory, or the like. The storage device 5 may be configured to be detachable from the diet suggestion device 100. The storage device 5 records various programs executed by the processor 1. The storage device 5 may also store machine learning models and learning data used in the health condition prediction process described later.

[0021] (Functional configuration) 3 is a block diagram showing the configuration of the meal suggestion device 100. Functionally, the meal suggestion device 100 includes an acquisition unit 101, a meal suggestion unit 102, a basis information generation unit 103, and a presentation unit 104.

[0022] The acquisition unit 101 acquires employee-related data 30, which is data related to the target employee 10. The acquisition unit 101 may acquire the employee-related data 30 from the storage device 5. Alternatively, the acquisition unit 101 may acquire the employee-related data 30 inputted via a device used by the employee. The employee-related data 30 is information related to the target employee 10. FIG. 4 shows an example of the employee-related data 30. The employee-related data 30 shown in FIG. 4 includes an employee ID (Identification), which is identification information of the target employee 10, a work start time, a work end time, a break start time, a break end time, a content of the task 1 performed by the target employee 10, a task 1 start time, a task 1 end time, a content of the task 2 performed by the target employee 10, a task 2 start time, and a task 2 end time. In this way, the employee-related data 30 may include data related to the employee's work schedule. In addition, the employee-related data 30 may further include not only a work schedule but also an actual work situation. Furthermore, the employee-related data 30 may include the behavioral history of the target employee 10 (such as activity level, number of steps, posture information, travel distance, commuting time, commuting route, etc.), or communication history with other employees or customers (such as conversation content, email content, chat content, reaction content, emotional data between the employee and the communication partner involved in the communication, etc.). The employee-related data 30 may also include information regarding the employee's stress. For example, the employee-related data 30 may include a stress value estimated based on commuting time. For example, if the commuting time is longer than usual, the employee is more likely to be stressed, and therefore the employee-related data 30 may include information indicating the employee's stress level.

[0023] The employee-related data 30 may include the results of the health checkup of the target employee 10 (height, weight, body fat percentage, body age, BMI (Body Mass Index), basal metabolism, and visceral fat level, etc.). The employee-related data 30 may also include weather, temperature, pedometer measurement information (date, number of steps, number of steps by time period, etc.), blood pressure measurement information (maximum blood pressure, minimum blood pressure, pulse rate, measurement time, etc.), weight measurement information (weight measurement time, etc.), vital data (pulse, respiration, body temperature, etc.), lifestyle information (mood, physical condition, diet, exercise, sleep, smoking status, drinking status, etc.), attribute information (nickname, sex, date of birth, age, family structure, etc. of the target employee 10), sleep-related data (measurement date, actual sleep time, time to fall asleep, awakening time / duration / number of times, quality of sleep, number of snoring, snoring level, etc.), etc. The employee-related data 30 may also include dietary intake information indicating the food consumed. For example, dietary intake information includes the content, calories, and time of meals (breakfast, lunch, dinner, etc.), the type and amount of drinks consumed, and the intake amount of each nutrient (protein, carbohydrate, lipid, vitamin, etc.).

[0024] The meal suggestion unit 102 determines a meal to be suggested to the target employee 10 using a meal selection model that has been machine-learned on the relationship between the employee-related data 30 and the suggested meal information 40. The employee-related data 30 used in the machine learning of the meal selection model may be collected on any target person. This target person may or may not include the target employee 10. The suggested meal information 40 may include information such as food name, ingredients, calorie information of the meal, nutritional information, serving place, and serving place address. The meal suggestion unit 102 determines a meal to be suggested to the target employee 10 using a meal selection model that has been machine-learned on the relationship between data to be considered in determining the proposed meal and the meal corresponding to the data. The data to be considered in determining the proposed meal may be included in the employee-related data 30. For example, the proposed meal may be determined taking into consideration the working hours, the work content performed, the age, and the nutrients of the most recently consumed meal. In this case, the meal suggestion unit 102 determines the meal to be suggested using a meal selection model that presents information indicating the meal to be suggested by inputting data indicating working hours, work performed, age, and nutrients in the most recently consumed meal (feature quantities described below).

[0025] The output of the meal selection model may indicate a specific meal (e.g., information indicating the name of a menu, the name of a dish, etc.), or may indicate conditions that the meal should satisfy (e.g., calories, ingredients, nutrition, price, etc.). In the latter case, the meal suggestion unit 102 may generate suggested meal information 40 indicating the output conditions, or may search for meals that satisfy the output conditions and generate suggested meal information 40 indicating the detected meals. In addition, the suggested "meal" may be anything that the target employee 10 can eat, and may be rephrased as, for example, a menu, a dish, a food, a meal plan, or food.

[0026] In this embodiment, an example will be described in which the meal suggestion unit 102 calculates the suggested meal information 40 as an output result of the meal selection model. The suggested meal information 40 is information on meals (food) suggested to the target employee 10. The form of the suggested meal information 40 is not particularly limited, and may include, for example, the name of the meal, the place of serving, calories, and nutrients (protein, carbohydrate, lipid, vitamin, etc.). The meal selection model is a model machine-learned using the learning data of the meal selection model, and may be stored in, for example, the storage device 5. The meal selection model may be generated by a known machine learning algorithm (for example, random forest, support vector machine, naive Bayes, or neural network, etc.). A plurality of meals may be suggested using a plurality of output values ​​of the model. The learning data of the meal selection model is data in which explanatory variables (feature values ​​described later) that are input to the meal selection model are associated with objective variables that are output from the meal selection model.

[0027] The objective variables of the meal selection model are used to generate the suggested meal information 40. The explanatory variables of the meal selection model are generated from the employee-related data 30. For example, the learning data may be created by associating, as the objective variables, information indicating meals proposed by a registered dietitian or the like with specialized knowledge with the explanatory variables generated from the employee-related data 30. As described above, the employee-related data 30 may include information indicating the degree of stress, and in this case, the information indicating the degree of stress may be used as the explanatory variable. When the information indicating the degree of stress is used as the explanatory variable, the learning data may be created by associating, as the objective variable, information indicating meals proposed by a social insurance labor consultant, an industrial counselor, or the like with specialized knowledge about labor with the explanatory variable.

[0028] The meal suggestion device 100 may have a function of generating learning data, i.e., a learning data generation unit. The learning data generation unit generates learning data by associating the proposed meal information 40 with the known employee-related data 30 acquired. The learning data generation unit may also generate learning data after performing preprocessing such as data cleansing, data integration, and data conversion on the employee-related data 30. The meal suggestion device 100 may also have a function of generating a meal selection model using the learning data, i.e., a meal selection learning unit. The meal selection learning unit generates a meal selection model by machine learning using the learning data. When the employee-related data 30 acquired by the acquisition unit 101 is input, the meal selection model generated using the learning data outputs the proposed meal information 40 based on the learning result.

[0029] The meal suggestion unit 102 may determine a meal based on a meal selection rule base without using a meal selection model. The meal selection rule base stores a rule for selecting a certain meal when the employee-related data 30 satisfies a certain condition. For example, the meal selection rule base may store a rule that "select a meal of X (kcal) or more" in association with a condition that "the work content is stocking shelves". For example, the meal selection rule base may store a rule that "select a sweet food such as dessert" in association with a condition that "the emotional data of the communication partner of the target employee 10 is anger". The timing when the meal suggestion unit 102 determines a meal is not particularly limited. For example, it may be at the end of a shift, at the start of a break, or when a task is changed.

[0030] The basis information generating unit 103 generates basis information 50 based on the contribution of the feature amount input to the meal selection model to the output value of the meal selection model. The feature amount is an explanatory variable of the meal selection model, and is either or both of data extracted from the employee-related data 30 and data obtained by processing the employee-related data 30. For example, the basis information generating unit 103 can use an XAI (eXplainable AI) algorithm to calculate the contribution of each feature amount to the output value of the meal selection model. In the following, an example will be described in which the basis information generating unit 103 uses LIME (Local Interpretable Model-agnostic Explanations), which is an XAI. The detailed operation of LIME will be described below. The basis information generating unit 103 may also use other XAI algorithms (for example, SHAP: Shapley Additive exPlanations, etc.) instead of LIME.

[0031] The LIME used by the basis information generating unit 103 will be described. LIME is a method for estimating the contribution of a feature to an output result from a change in the feature and a change in the output result of a model. The calculation method is expressed by the following formula (1).

[0032]

number

[0033]

number

[0034] Equation (2) is the output result for the value obtained by randomly changing the feature quantity input to the meal selection model, and y is the output result for the value of the feature quantity input to the meal selection model. i is the feature vector x i It is a parameter (variable) that indicates the contribution of to the output result.

[0035] ε i (n) is each feature xi The difference between the features and the degree of change in the feature when the feature is randomly changed. N is the number of random samples, and d is the number of features. The basis information generating unit 103 calculates a i Derive and derive a i is the contribution of each feature to the output result. In other words, the basis information generating unit 103 calculates the output result of the meal selection model and i Minimize the sum of squares error between the output result of the approximate model and the food choice model approximated using a i By solving the optimization problem to find i Calculate.

[0036] Furthermore, the basis information generating unit 103 may select a feature amount to be presented as the basis information 50 based on the calculated contribution degree. For example, the basis information generating unit 103 may present, as the basis information 50, a feature amount with a relatively high contribution degree, or a feature amount with a contribution degree equal to or greater than a predetermined threshold, among the feature amounts whose contribution degrees are calculated using the above-mentioned formula (1). The basis information 50 may indicate the feature amount and its contribution degree.

[0037] When the meal suggestion unit 102 selects a meal based on the meal selection rule base, the basis information generation unit 103 may generate basis information 50 indicating a condition that has been satisfied in the meal selection rule base. For example, when a "meal of X (kcal) or more" is selected because the condition "the task was stocking" has been satisfied, the basis information generation unit 103 may generate basis information 50 indicating that "the task was stocking".

[0038] The presentation unit 104 presents the proposed meal information 40 selected by the meal proposal unit 102 and the basis information 50 calculated by the basis information generation unit 103 to the terminal device 20. FIG. 5 shows an example of presentation to the terminal device 20 in the first embodiment. As shown in FIG. 5, the proposed meal information 40 may be, for example, information indicating the meal serving location in addition to the food name. The proposed meal information 40 may also include related information such as ingredients, calories, and nutritional information of the proposed meal. The proposed meal information 40 may also include the address of the serving location. As shown in FIG. 5, the basis information 50 may also include feature amounts and their contribution rates. As shown in FIG. 5, the basis information 50 may also be a graph showing the contribution rate of each feature amount. This makes it possible to propose a meal to the target employee 10 after clarifying the reason for proposal. The proposed meal information 40 and the basis information 50 presented by the presentation unit 104 may be viewed not only by the target employee 10 but also by other users. The other user may be, for example, the superior of the target employee 10 or an industrial doctor at the workplace of the target employee 10. The presentation form of the suggested meal information 40 and the basis information 50 is not limited to display, and may be presented in the form of voice, printout, or the like.

[0039] (Information Processing) 6 is a flowchart of processing by the meal suggestion device 100 of the first embodiment. This processing is realized by the processor 1 shown in FIG. 2 executing a program prepared in advance and operating as each element shown in FIG.

[0040] First, the acquisition unit 101 acquires the employee-related data 30 (step S11). Next, the meal suggestion unit 102 uses a meal selection model that has been machine-learned to learn the relationship between the employee-related data 30 and the suggested meal information 40 to determine a meal to be proposed to the target employee 10 from the employee-related data 30 acquired in S11 (step S12). Next, the basis information generation unit 103 generates basis information 50 for the prediction result of each feature amount used in the prediction based on the meal selection model used in the prediction in S12 and the employee-related data 30 acquired by the acquisition unit 101 in S11 (step S13). The contribution degree to the prediction result for calculating the basis information 50 can be calculated by using, for example, the XAI algorithm, as described above. Next, the presentation unit 104 presents the suggested meal information 40 and the basis information 50 to the terminal device 20 (step S14). This ends the process.

[0041] As described above, in the first embodiment, it is possible to propose a meal to the target employee 10 after clarifying the reason for the selection.

[0042] (Variation 1) In the above-mentioned first embodiment, the meal suggestion device 100 includes an acquisition unit 101, a meal suggestion unit 102, a basis information generation unit 103, and a presentation unit 104, and determines a meal to be suggested to the target employee 10. As a further modification, an incentive information determination unit 105 may be provided. FIG. 7 is a block diagram showing the configuration of the meal suggestion device 100 according to the first modification.

[0043] The incentive information determination unit 105 determines incentive information 60 related to the suggested meal information 40. The incentive information 60 is information that serves as a factor for motivating the person to whom the information is presented to take action. For example, the incentive information 60 may be in the form of a coupon, points, a gift certificate, electronic money, a product exchange voucher, or the like related to the suggested meal information 40. The incentive information 60 can encourage the target employee 10 to actually purchase the suggested meal.

[0044] The incentive information determination unit 105 may determine the incentive information 60 associated with the suggested meal information 40 based on an incentive information database stored in the storage device 5. The incentive information database may store, for example, the type of incentive, the content of the incentive, the target food, and the determination conditions. The incentive information database may also include expiration date (best before) information and inventory information of the target food, and the incentive information determination unit 105 may determine the incentive information 60 related to the suggested meal information 40 as the incentive information 60 related to the food that is approaching its expiration date (best before).

[0045] For example, if the target employee 10 works in a supermarket, the incentive information database may include a product database used in the supermarket. The product database includes product names, product types, best-before (expiration) dates, stock quantities, and stock stores. The incentive information database stores incentive information 60 in combination with products whose expiry dates are within X hours and whose product type is prepared food.

[0046] The presentation unit 104 presents the proposed meal information 40 determined by the meal proposal unit 102, the basis information 50 calculated by the basis information generation unit 103, and the incentive information 60 determined by the incentive information determination unit 105 to the terminal device 20. The presentation of the proposed meal information 40 and the basis information 50 is the same as in the above embodiment. The incentive information 60 may be, for example, a coupon that the target employee 10 can apply when purchasing the proposed meal information 40, or points that are given to the target employee 10 when purchasing the proposed meal information 40. The incentive information 60 may also include coupons, points, gift certificates, electronic money, product exchange coupons, etc. related to the proposed meal information 40. In this way, the meal proposal device 100 of the first modified example can propose a meal to the target employee 10 together with the incentive information 60, while clearly indicating the reason for selection. The presentation mode of the incentive information 60 is not limited to display, and may be presented in the form of voice, print, etc.

[0047] Fig. 8 shows an example of presentation to terminal device 20 in variant example 1. The presentation example in Fig. 8 differs from the presentation example in Fig. 5 in that incentive information 60 is also displayed. Incentive information 60 shown in Fig. 8 displays a coupon for receiving a discount on the purchase price of "Lunch Box A" shown in suggested meal information 40, and a coupon for earning points when purchasing "Lunch Box A". By presenting this coupon when purchasing "Lunch Box A", for example, target employee 10 can enjoy the benefits of a discount or points.

[0048] 9 is a flowchart of processing by the meal recommendation device 100 of Modification 1. This processing is realized by the processor 1 shown in FIG. 2 executing a prepared program and operating as each element shown in FIG.

[0049] First, the acquisition unit 101 acquires the employee-related data 30 (step S21). Next, the meal suggestion unit 102 uses a meal selection model that has been machine-learned to learn the relationship between the employee-related data 30 and the proposed meal information 40 to determine a meal to be proposed to the target employee 10 from the employee-related data 30 acquired in S21 (step S22). Next, the basis information generation unit 103 generates basis information 50 for the prediction result of each feature amount used in the prediction based on the meal selection model used in the prediction in S22 and the employee-related data 30 acquired by the acquisition unit 101 in S21 (step S23). The contribution to the prediction result for calculating the basis information 50 can be calculated by using, for example, the XAI algorithm, as described above. Next, the incentive information determination unit 105 determines incentive information 60 related to the proposed meal information 40 determined in S23 (step S24). Next, the presentation unit 104 presents the proposed meal information 40, the basis information 50, and the incentive information 60 to the terminal device 20 (step S25). This completes the process.

[0050] (Variation 2) In addition, in the first embodiment described above, the suggested meal information 40 may be determined based on the location information of the employee. In this case, the acquisition unit 101 acquires location information indicating the location of the employee as the employee-related data 30. Furthermore, the meal information that is a candidate for the suggested meal information 40 includes the address of the meal providing place. FIG. 10 shows an example of the meal information. The meal information may include the food name, calories, the name of the providing place, and the like, in addition to the address of the providing place.

[0051] The meal suggestion unit 102 determines the meal to be suggested to the target employee 10 using a meal selection model that has been machine-learned on the relationship between the employee-related data 30 and the suggested meal information 40. In the second modification, by including the employee's location information and the address of the meal provision location as learning data for the meal selection model, meals can be suggested that take into account the distance and travel time between the employee and the meal provision location. For example, as learning data, meal information (objective variable) that makes the employee's break time included in the employee-related data 30 longer than the travel time calculated from the address of the meal provision location and the employee's location information is combined with the employee-related data 30, which is an explanatory variable. This makes it possible to suggest meals in accordance with the location information of the target employee 10.

[0052] Furthermore, the meal proposal unit 102 may determine, as the meal to be proposed to the target employee 10, a meal provided at a place within a predetermined range from the location of the target employee 10, among the meals determined using the meal selection model described in the first embodiment. In this case, the meal proposal unit 102 identifies the location of the target employee 10 from the location information of the target employee 10, and identifies the location where the meal is provided, determined using the meal selection model. Then, if the identified location is within a predetermined range from the location of the target employee 10, the meal proposal unit 102 determines the meal as the meal to be proposed to the target employee 10. Note that the predetermined range may be determined by the straight-line distance from the location to the location, the distance of the route from the location to the location, or the time required to travel from the location to the location. Furthermore, if the employee-related data 30 includes data indicating the employee's break time or the average time required for a meal, the meal proposal unit 102 may determine the above-mentioned predetermined range based on the break time or the average time required for a meal.

[0053] The presentation unit 104 presents the suggested meal information 40 and the basis information 50 to the terminal device 20. FIG. 11 shows an example of presentation to the terminal device 20. The presentation example of FIG. 11 differs from the presentation example of FIG. 5 in the content of the suggested meal information 40. As in the example of FIG. 11, the presentation unit 104 may display the meal provision points and the employee's location information on a map as the suggested meal information 40. The employee's location information may be updated at regular intervals. The presentation unit 104 may display information (e.g., icons) indicating the meal provision points and the employee's locations on a map displayed by an existing map application as the suggested meal information 40. Note that the presentation mode of the suggested meal information 40 and the basis information 50 is not limited to display, and may be presented in the form of audio, printing, or the like.

[0054] (Variation 3) In the above-mentioned first embodiment, the meal suggestion device 100 includes an acquisition unit 101, a meal suggestion unit 102, a basis information generation unit 103, and a presentation unit 104, and determines a meal to be suggested to the target employee 10. As a further modification, an instruction unit 107 may be provided. FIG. 12 is a block diagram showing the configuration of the meal suggestion device 100 according to the third modification.

[0055] The instruction unit 107 provides the necessary support for the target employee 10 to eat the suggested meal. For example, the instruction unit 107 may instruct the target employee 10 to deliver the suggested meal to the workplace of the target employee 10. Also, for example, the instruction unit 107 may instruct the target employee 10 to deliver the suggested meal to a refrigerated locker at a train station by X o'clock.

[0056] Also, for example, the instruction unit 107 may place an order for the suggested meal. Note that, when the target employee 10 visits a place where meals are provided, the instruction unit 107 may place an order or make a reservation so that the meal is provided at the scheduled arrival time. Also, the instruction unit 107 may place an order for a take-out meal.

[0057] The instruction unit 107 may transmit instructions to the store that provides the meal. The instruction unit 107 may also transmit the instruction contents to another delivery and ordering system. The delivery and ordering system, for example, accepts a delivery order to the store and places an order with a delivery company. This makes it easy for the target employee 10 to actually consume the proposed meal.

[0058] <Second embodiment> Next, a second embodiment of the present disclosure will be described. The second embodiment differs from the first embodiment in the following respects. The meal suggestion device 100 of the second embodiment includes a route optimization unit 106 that optimizes a route that passes through suggested meal providing locations, and the presentation unit 104 presents optimal route information in addition to the suggested meal information 40 and the evidence information 50. Other than these respects, the second embodiment is similar to the first embodiment. Specifically, the overall configuration of the meal suggestion system 1000 and the hardware configuration of the meal suggestion device 100 of the second embodiment are similar to the overall configuration shown in FIG. 1 and the hardware configuration shown in FIG. 2 of the first embodiment, and therefore description thereof will be omitted.

[0059] (Functional configuration) Fig. 12 is a block diagram showing the functional configuration of the meal suggestion device 100. As shown in the figure, the meal suggestion device 100 of the second embodiment includes an acquisition unit 101, a meal suggestion unit 102, a basis information generation unit 103, a presentation unit 104, and a route optimization unit 106. Note that the acquisition unit 101, the meal suggestion unit 102, and the basis information generation unit 103 are similar to those of the first embodiment, and therefore description thereof will be omitted.

[0060] The route optimization unit 106 optimizes a route that passes through locations where suggested meals are provided, based on a mathematical optimization calculation method using constraint conditions that include at least a portion of the suggested meal information 40 and the employee-related data 30 as conditions.

[0061] The mathematical optimization calculation method is a method of defining a problem as a mathematical formula and finding the values ​​of variables that minimize or maximize an objective function while satisfying constraints. For example, known algorithms such as dynamic programming, greedy algorithms, and approximation algorithms may be used.

[0062] First, the route optimization unit 106 receives map data and constraint conditions. The map data may include road data, building data, topographical data, traffic data, facility information, and noise indices for facilities and roads.

[0063] The constraint conditions are conditions that the optimal route information 70 must satisfy, and include at least a portion of the suggested meal information 40 and the employee-related data 30. For example, the constraint conditions include the following conditions: "The target employee 10 passes through a meal service location," "The target employee 10 walks for more than X minutes," "The route takes less than the target employee 10's break time," "The route passes through a park," and "The route goes along a slope." The constraint conditions can be set and changed by the target employee 10.

[0064] Next, the route optimization unit 106 generates an objective function. The objective function is a function to be maximized or minimized in a mathematical optimization calculation method. Evaluation items included in the objective function are, for example, "travel distance of the target employee 10," "travel time of the target employee 10," "total noise index of points and roads where the target employee 10 stops," etc. A weighting value may be set for each evaluation item. The evaluation items and weighting values ​​may change based on the employee-related data 30. The evaluation items and weighting can be set and changed by the target employee 10.

[0065] The presentation unit 104 presents the proposed meal information 40 determined by the meal proposal unit 102, and the basis information 50 and optimal route information 70 calculated by the basis information generation unit 103 to the terminal device 20. FIG. 14 shows an example of presentation to the terminal device 20 in the second embodiment. The output of the proposed meal information 40 and the basis information 50 is the same as in the first embodiment. The optimal route information 70 may be, for example, a line (e.g., a dotted line) indicating the optimal route displayed on a map. The optimal route information 70 may also be displayed on a map output as the proposed meal information 40. This allows the target employee 10 to be proposed a meal along an optimal route to a service location that provides the proposed menu. Note that the presentation mode of the proposed meal information 40, the basis information 50, and the optimal route information 70 is not limited to display, and may be output in the form of voice, printout, or the like.

[0066] (Information Processing) 15 is a flowchart of processing by the meal suggestion device 100 of the second embodiment. This processing is realized by the processor 1 shown in FIG. 2 executing a program prepared in advance and operating as each element shown in FIG.

[0067] First, the acquisition unit 101 acquires the employee-related data 30 (step S31). Next, the meal suggestion unit 102 uses a meal selection model that has been machine-learned to learn the relationship between the employee-related data 30 and the proposed meal information 40 to determine a meal to be proposed to the target employee 10 from the employee-related data 30 acquired in S31 (step S32). Next, the basis information generation unit 103 generates basis information 50 for the prediction result of each feature amount used in the prediction based on the meal selection model used in the prediction in S32 and the employee-related data 30 acquired by the acquisition unit 101 in S31 (step S33). The contribution to the prediction result for calculating the basis information 50 can be calculated by using, for example, the XAI algorithm, as described above. Next, the route optimization unit 106 optimizes the route that passes through the places where the proposed meal is provided, based on a mathematical optimization calculation method (step S34). Next, the presentation unit 104 outputs the proposed meal information 40, the basis information 50, and the optimal route information 70 to the terminal device 20 (step S35). This completes the process.

[0068] As described above, in the second embodiment, meals can be proposed to the target employee 10 after clarifying the reasons for the selection along with the optimal route.

[0069] <Third embodiment> Next, a third embodiment of the present disclosure will be described. The third embodiment differs from the first embodiment in the following two points. The first point is that the meal suggestion device 100 of the third embodiment includes a sentence generation unit 108 in addition to the units included in the meal suggestion device 100 of the first embodiment. The sentence generation unit 108 generates a sentence (which can also be called a text) using the suggested meal information 40, the evidence information 50, and a language model. The second point is that the presentation unit 104 outputs the sentence generated by the sentence generation unit 108. Other than these points, the third embodiment is similar to the first embodiment. Specifically, the hardware configuration of the meal suggestion device 100 of the third embodiment is similar to the hardware configuration of the first embodiment shown in FIG. 2, and therefore a description thereof will be omitted.

[0070] (Functional configuration) 16 is a block diagram showing the functional configuration of the meal suggestion device 100. As shown in the figure, the meal suggestion device 100 of the third embodiment includes an acquisition unit 101, a meal suggestion unit 102, a basis information generation unit 103, a presentation unit 104, and a sentence generation unit 108. Note that the acquisition unit 101, the meal suggestion unit 102, and the basis information generation unit 103 are the same as those in the first embodiment, and therefore description thereof will be omitted.

[0071] The sentence generation unit 108 generates a sentence by using the suggested meal information 40, the evidence information 50, and the language model. The sentence generation unit 108 transmits the generated sentence to the presentation unit 104. Then, the presentation unit 104 outputs the sentence generated by the sentence generation unit 108.

[0072] (About the language model) Here, the language model used by the sentence generation unit 108 in this embodiment will be described. The language model is a model that learns the relationship between words in a sentence and generates related strings related to a target string from the target string. By using a language model that has learned sentences and paragraphs in various contexts, it is possible to generate related strings with appropriate content related to the target string.

[0073] For example, a case where a language model is used in question answering will be described. The language model receives an input of a question such as "What kind of country is Japan?" as a target string. The language model generates a related string such as "Japan is an island country in the northern hemisphere..." as an answer to the question.

[0074] The method of training the language model is not particularly limited, but as one example, the language model may be trained to output at least one sentence that includes an input character string.

[0075] As a specific example, the language model is a GPT (Generative Pre-Training) model that outputs a sentence including an input string by predicting a string that is likely to follow the input string. In addition, for example, T5 (Text-to-text Transfer Transformer), BERT (Bidirectional Encoder Representations from Transformers), RoBERTa (Robstly optimized BERT approach), and ELECTRA (Efficiently Learning an Encoder that Classifies Token Replacements Accurately) are also language models.

[0076] This concludes the description of the language model. The following describes an overview of how the sentence generation unit 108 generates a sentence using the language model. First, the sentence generation unit 108 acquires an input sentence (prompt). The prompt is a sentence that commands the generation of a sentence that suggests a meal to an employee, and includes either or both of the suggested meal information 40 and the basis information 50.

[0077] The language model may be stored in any storage device accessible to the meal suggestion device 100, such as the storage device 5. The prompt may be stored in advance in the storage device 5 or the like, or may be generated by the sentence generation unit 108. For example, the sentence generation unit 108 can generate a prompt by inputting various information included in the suggested meal information 40 and the basis information 50 into a template generated in advance.

[0078] For example, the prompt may read, "Suggest a meal to the employee along with the evidence information. The suggested meal is the "Stamina Bowl" sold at "Genki Bento". The employee will be performing the task of "stocking shelves" in the morning, and the stamina bowl is being suggested because the employee is expected to be physically tired. Please output a comment for this employee." Such a prompt can be generated from a template in which the parts "Genki Bento", "Stamina Bowl", and "Stocking shelves" are left blank. In this case, the sentence generation unit 108 can generate the prompt by inputting the information on "suggested meal information" and "name of place of provision" shown in the suggested meal information 40 and the "task" shown in the employee-related data 30 into the template. The user may also edit the prompt using the input / output interface 2.

[0079] When the sentence generation unit 108 receives the input of the prompt, it uses a language model to generate a comment 80 suggesting a meal to the employee. The sentence generation unit 108 transmits the generated comment 80 to the presentation unit 104. The sentence generation unit 108 may change the output of the language model based on conditions (such as the size of the display area and the number of characters) specified by the user. For example, the sentence generation unit 108 calculates the number of characters that will fit within the display area based on the size of the display area, and generates a comment 80 that is equal to or less than that number of characters. For example, the sentence generation unit 108 can use a prompt in which a limit on the number of characters is described to generate a comment 80 that satisfies the limit.

[0080] The presentation unit 104 of the third embodiment differs from the presentation unit 104 of the first embodiment in that it outputs a comment 80 in addition to the suggested meal information 40 and the basis information 50. As shown in Fig. 17, the presentation unit 104 outputs the suggested meal information 40 determined by the meal suggestion unit 102, the basis information 50 generated by the basis information generation unit 103, and the comment 80 generated by the sentence generation unit 108 to the terminal device 20. Note that these pieces of information do not necessarily need to be displayed simultaneously.

[0081] As described above, the meal suggestion device 100 of the third embodiment has a feature not found in the meal suggestion device 100 of the first and second embodiments, that is, it generates a sentence using the suggested meal information 40, the grounds information 50, and a language model. As a result, the meal suggestion device 100 of the third embodiment can suggest a meal to the target employee 10 after clarifying the reason for selection along with the grounds and the comment 80 explaining the contents of the suggested meal.

[0082] Although the present invention has been described above with reference to the embodiments, the present invention is not limited to the above embodiments. The configuration and details of the present invention may be modified in various ways that can be understood by a person skilled in the art within the scope of the present invention. The present disclosure may include embodiments in which the matters described in this specification are appropriately combined or replaced as necessary. For example, matters described using a specific embodiment (including drawings) may be applied to other embodiments to the extent that no contradiction occurs. For example, although multiple operations are described in order in the form of a flowchart, the order of description does not limit the order in which the multiple operations are performed. Therefore, when implementing each embodiment, the order of the multiple operations can be changed to the extent that the content is not affected. In addition, the processing of each step shown in each flowchart can be executed by one processor, or can be executed by multiple processors in a shared manner. In other words, the execution entity of the meal suggestion method according to each of the above-mentioned embodiments may be one processor or multiple processors.

[0083] A part or all of the above-described embodiments can be described as follows: However, a part or all of the above-described embodiments is not limited to the following.

[0084] (Appendix 1) an acquisition unit that acquires subject-related data related to a subject to whom a meal is to be proposed; a meal suggestion unit that uses the subject-related data to determine a meal to be proposed to the subject; a basis information generation unit that generates basis information indicating a basis for proposing the meal; and a presentation unit that presents the proposed meal and the basis information. A meal suggestion device comprising:

[0085] (Appendix 2) 2. The meal recommendation device according to claim 1, further comprising an incentive information determination unit that determines incentive information indicating an incentive for the subject to select the recommended meal.

[0086] (Appendix 3) The meal recommendation device described in Appendix 1 or 2, wherein the meal recommendation unit determines meals provided at locations within a predetermined range from the subject's location information as meals to be recommended to the subject, and the presentation unit presents a map showing the subject's location information and the locations where the meals are provided.

[0087] (Appendix 4) The meal recommendation device according to claim 3, further comprising a route optimization unit that optimizes a route from a location of the subject to a location where the meal is provided, using constraint conditions including at least a portion of the meal and the subject-related data as conditions, and the presentation unit displays the route optimized by the route optimization unit on the map.

[0088] (Appendix 5) A meal suggestion method, comprising: at least one processor acquiring subject-related data related to a subject to whom a meal is to be suggested; using the subject-related data to determine a meal to suggest to the subject; generating evidence information indicating a basis for suggesting the meal; presenting the suggested meal; and presenting the evidence information.

[0089] (Appendix 6) A meal recommendation program that causes a computer to execute the following processes: a process of acquiring subject-related data related to a subject to whom a meal is to be recommended; a process of determining a meal to be recommended to the subject using the subject-related data; a process of generating evidence information indicating the basis for recommending the meal; a process of presenting the recommended meal; and a process of presenting the evidence information.

[0090] The embodiments of Supplementary Note 5 and Supplementary Note 6 can be expanded to the embodiments of Supplementary Note 2 to Supplementary Note 4, similarly to Supplementary Note 1. [Explanation of symbols]

[0091] 1 processor 2 Input / Output Interface 3 ROM 4 RAM 5 Storage device 6 Bus 10 Eligible Employees 20 Terminal Equipment 30 Employee-related data 40 Suggested Meal Information 50 Basis information 60 Incentive Information 70 Best Route Information 80 Comments 100 Meal suggestion device 101 Acquisition Department 102 Meal Suggestions Department 103 Basis Information Generation Unit 104 Presentation section 105 Incentive Information Determination Unit 106 Route Optimization Department 107 Instruction section 108 Document Generation Department

Claims

1. An acquisition unit that acquires subject-related data related to a subject for whom a meal is to be proposed; a meal suggestion unit that uses the subject-related data to determine a meal to be suggested to the subject; a basis information generating unit that generates basis information indicating a basis for suggesting the meal; A presentation unit that presents the recommended meal and the basis information; A meal suggestion device comprising:

2. an incentive information determination unit that determines incentive information indicating an incentive for the subject to select the proposed meal; The meal suggestion device according to claim 1 .

3. The meal suggestion unit determines, as a meal to be suggested to the subject, a meal provided at a place within a predetermined range from the location information of the subject, The meal suggestion device according to claim 1 , wherein the presentation unit presents a map showing location information of the subject and a location where the meal is provided.

4. a route optimization unit that optimizes a route from a location of the subject to a place where the meal is provided, using constraint conditions that include at least a part of the meal and the subject-related data as conditions; The meal recommendation device according to claim 3 , wherein the presentation unit causes the route optimized by the route optimization unit to be displayed on the map.

5. The meal suggestion unit, said subject-related data, The relationship between the subject-related data and the meals corresponding to the subject-related data is input into a machine learning model to determine meals to be recommended to the subject. The meal suggestion device according to claim 1 .

6. At least one processor Obtaining subject-related data relating to a subject for which a meal recommendation is to be made; using the subject-related data to determine a suggested meal for the subject; generating evidence information indicating evidence for suggesting the meal; presenting said suggested meals; and presenting the evidence information.

7. obtaining subject-related data relating to the subject for which meal suggestions are made; using the subject-related data to determine a meal to be recommended to the subject; A process of generating basis information indicating a basis for suggesting the meal; presenting said meal suggestions; A process of presenting the grounds information; A meal recommendation program that causes a computer to execute the above.

Images