A learning support device, a learning support method, a learning device, and a method for generating an objective function.

Abstract

Recommended courses will be determined based on the student's current course selection status. [Solution] The course registration support device includes a data acquisition unit that acquires state data indicating the status of a subject's course selection, and a recommended course determination unit that determines which courses to recommend to the subject based on the results of an optimization calculation using an objective function generated by inverse reinforcement learning using training data indicating the status of course selection and the selected courses corresponding to that status, and the state data acquired by the data acquisition unit.

Description

[[Technical Field]]

[0001] The present disclosure relates to a course registration support apparatus, a course registration support method, a learning apparatus, a method for generating an objective function, and the like. [[Background Art]]

[0002] Technologies for supporting students and the like to appropriately register for courses are known. For example, Patent Document 1 below discloses a curriculum analysis and creation support system that selects and outputs courses that allow acquisition of capabilities related to student needs as recommended courses based on student information on student needs and course information on capabilities acquirable through courses. More specifically, in the curriculum analysis and creation support system described in Patent Document 1, each student and each course are represented by coordinates, and a predetermined number of courses that are close in distance from the student's coordinates are identified as recommended courses for the student. [[Prior Art Documents]] [[Patent Documents]]

[0003] [[Patent Document 1]] Japanese Unexamined Patent Application Publication No. 2006-139130 [[Summary of the Invention]] [[Problem to be Solved by the Invention]]

[0004] The curriculum analysis and creation support system of Patent Document 1 has room for improvement in that it cannot consider various states related to course selection when selecting recommended courses. For example, a plurality of recommended courses output by the curriculum analysis and creation support system of Patent Document 1 can include courses that cannot be taken at the same time due to scheduling conflicts, courses that have already been taken, and the like. This is because the curriculum analysis and creation support system of Patent Document 1 selects recommended courses only based on the criterion that the coordinate distance is close, without considering whether courses can be selected simultaneously or the status of courses already taken by each student.

[0005] This disclosure has been made in view of the above-mentioned issues, and one exemplary purpose is to provide a technique for determining which courses to recommend to a subject, taking into account the subject's status regarding course selection. [Means for solving the problem]

[0006] A course support device according to one aspect of the present invention comprises: a data acquisition means for acquiring state data indicating the state of a subject's course selection; and a recommended course determination means for determining courses to recommend to the subject based on the results of an optimization calculation using an objective function generated by inverse reinforcement learning using training data indicating the state of course selection and the selected courses corresponding to that state, and the state data acquired by the data acquisition means.

[0007] A learning device according to one aspect of the present invention comprises: data acquisition means for acquiring training data indicating a state regarding course selection and the courses selected according to that state; and learning means for generating an objective function for determining courses to recommend to a subject, according to the subject's state regarding course selection, by performing inverse reinforcement learning using the training data.

[0008] In a course enrollment support method according to one aspect of the present invention, at least one processor performs a data acquisition process to acquire state data indicating the state of a subject's course selection, and a recommended course determination process to determine which courses to recommend the subject to take based on the results of an optimization calculation using an objective function generated by inverse reinforcement learning using training data indicating the state of course selection and the selected courses corresponding to that state, and the state data acquired in the data acquisition process.

[0009] In a method for generating an objective function according to one aspect of the present invention, at least one processor performs a data acquisition process to acquire training data indicating a state regarding course selection and the courses selected according to that state, and a learning process to generate an objective function for determining which courses to recommend to the subject, according to the state regarding course selection of the subject, by performing inverse reinforcement learning using the training data. [Effects of the Invention]

[0010] According to an illustrative aspect of this disclosure, one illustrative effect is that it becomes possible to determine which courses to recommend to a subject, taking into account the subject's status regarding course selection. [Brief explanation of the drawing]

[0011] [Figure 1] This block diagram shows the configuration of the course support system related to this disclosure. [Figure 2] This is a flowchart showing the flow of the method for generating the objective function related to this disclosure. [Figure 3] This flowchart shows the flow of the course support methods related to this disclosure. [Figure 4] This is a block diagram showing the configuration of other learning support devices related to this disclosure. [Figure 5] This figure shows an example of generating an objective function. [Figure 6] This diagram shows an example of how recommended courses are determined. [Figure 7] Figure 4 is a flowchart illustrating an example of the process performed by the learning support system. [Figure 8] This is a block diagram showing the configuration of a computer that functions as an information processing device related to this disclosure. [Modes for carrying out the invention]

[0012] The following are examples of embodiments of the present invention. However, the present invention is not limited to the exemplary embodiments shown below, and various modifications are possible within the scope of the claims. For example, embodiments obtained by appropriately combining some or all of the technologies (things or methods) employed in each of the exemplary embodiments shown below may also be included in the scope of the present invention. Furthermore, embodiments obtained by appropriately omitting some of the technologies employed in each of the exemplary embodiments shown below may also be included in the scope of the present invention. In addition, the effects mentioned in each of the exemplary embodiments shown below are examples of effects that can be expected in that exemplary embodiment and do not define the scope of the present invention. That is, embodiments that do not produce the effects mentioned in each of the exemplary embodiments shown below may also be included in the scope of the present invention.

[0013] [First Exemplary Embodiment] A first exemplary embodiment, which is an example of an embodiment of the present invention, will be described in detail with reference to the drawings. This exemplary embodiment is the basic form for each of the exemplary embodiments described later. The scope of application of each technology adopted in this exemplary embodiment is not limited to this exemplary embodiment. That is, each technology adopted in this exemplary embodiment can also be adopted in other exemplary embodiments included in this disclosure, to the extent that no particular technical problems occur. Furthermore, each technology shown in the drawings referenced to explain this exemplary embodiment can also be adopted in other exemplary embodiments included in this disclosure, to the extent that no particular technical problems occur.

[0014] (Regarding the Course Registration Support System 1) The configuration of the course support system 1 according to this exemplary embodiment will be described with reference to Figure 1. Figure 1 is a block diagram showing the configuration of the course support system 1. The course support system 1 is a system for supporting the selection of courses to be taken by a target student, and includes a learning device 11 and a course support device 12 as shown in the figure.

[0015] (Regarding the learning device 11) The learning device 11 is a device that generates an objective function used to support a subject's selection of courses to take. As shown in FIG. 1, the learning device 11 comprises a data acquisition unit 111 and a learning unit 112. Details of the objective function will be described below.

[0016] The data acquisition unit 111 acquires training data indicating a state related to course selection and a selected course corresponding to the state. The required number of pieces of training data are acquired for inverse reinforcement learning described later. That is, the data acquisition unit 111 acquires a training dataset including training data required for inverse reinforcement learning. The method for acquiring state data is arbitrary. For example, the data acquisition unit 111 may acquire training data input to the learning device 11, or may acquire training data from another device.

[0017] Here, the courses to take refer to subjects that are to be taken by students or the like at educational institutions such as universities, high schools, or vocational schools. Here, the term "subject" can also be rephrased as "course subject", "class", "lecture", "seminar", "course", or the like. Furthermore, the term "taking a course" can also be rephrased as "attending a class", "selecting a subject", or the like.

[0018] Furthermore, the "state related to course selection" is a state that influenced or may have influenced the selection of a course to take, and can also be rephrased as "situation at the time of course selection", "conditions for course selection", or the like. For example, at the time of selecting a course to take, selectable subjects and the student who selected the course can be said to constitute a state related to course selection.

[0019] Furthermore, elective courses corresponding to a state regarding course selection are elective courses selected under a certain state regarding course selection. These elective courses only need to include at least one course. In other words, in a single training dataset, one or more elective courses are associated with a state regarding course selection. Note that the state regarding course selection does not necessarily have to be a state that actually existed, and the elective courses corresponding to that state do not necessarily have to be courses that were actually selected.

[0020] The learning unit 112 generates an objective function for determining which subjects to recommend to the subject, based on the subject's state regarding subject selection, by performing inverse reinforcement learning using the training data acquired by the data acquisition unit 111. Inverse reinforcement learning is a method that estimates the objective function using training data that shows the optimal actions in various states, whereas reinforcement learning is a method that learns the objective function for an agent to select the optimal action in an environment. The objective function generated by inverse reinforcement learning is a function for evaluating various actions in various states, and can also be called a "reward function," etc.

[0021] As described above, the learning device 11 according to this exemplary embodiment is configured to include a data acquisition unit 111 that acquires training data indicating the state of course selection and the courses selected according to that state, and a learning unit 112 that generates an objective function for determining which courses to recommend to the subject, according to the state of the subject selection, by performing inverse reinforcement learning using the training data acquired by the data acquisition unit 111.

[0022] According to the above configuration, an objective function is generated to determine which courses to recommend to the subject, based on the subject's current course selection status. By using this objective function, it becomes possible to determine which courses to recommend to the subject, based on the subject's current course selection status. Therefore, according to the above configuration, the effect is obtained that it becomes possible to determine which courses to recommend to the subject, taking into account the subject's current course selection status.

[0023] (Regarding the course registration support device 12) The course registration support device 12 is a device that assists students in selecting courses to take. As shown in Figure 1, the course registration support device 12 includes a data acquisition unit 121 and a recommended course determination unit 122.

[0024] The data acquisition unit 121 acquires status data indicating the status of the subject's course selection. The method of acquiring the status data is arbitrary. For example, the data acquisition unit 121 may acquire status data input to the course support device 12, or it may acquire status data from another device.

[0025] The "status data" mentioned above only needs to indicate conditions that may influence the subject's choice of courses. For example, the subjects the subject can choose from, the subject's personality, areas of interest, subjects the subject excels at / doesn't excel at, the subject's desired career path, and the qualifications the subject wants to obtain can all be considered conditions related to course selection. "Status data" can also be rephrased as "situation data" that indicates the circumstances at the time of the subject's course selection, or "condition data" that indicates the conditions for the subject's course selection.

[0026] The recommended course determination unit 122 determines the courses that the target individual should take (hereinafter sometimes referred to as recommended courses) based on the state data acquired by the data acquisition unit 121. More specifically, the recommended course determination unit 122 determines the recommended courses based on the results of an optimization calculation using the objective function generated by inverse reinforcement learning and the state data acquired by the data acquisition unit 121. These recommended courses only need to include at least one other course. In other words, the recommended course determination unit 122 determines one or more courses as recommended courses. When multiple courses are determined as recommended courses, it can also be said that the recommended course determination unit 122 determines the combination of courses that are recommended for study.

[0027] As mentioned above, inverse reinforcement learning is a method for estimating an objective function using training data that shows the optimal action in various states. The inverse reinforcement learning used to generate the objective function used in the optimization calculation above uses training data that shows the selected subjects according to the state regarding course selection. This objective function may be generated by the learning unit 112 of the learning device 11, or it may be generated by another device. Furthermore, the optimization calculation may be performed by the recommended course determination unit 122, or by providing another block in the course support device 12 and having that block perform the calculation, or by another device.

[0028] Furthermore, determining recommended courses based on the results of optimization calculations means that the recommended courses are determined by at least utilizing the results of the optimization calculations. For example, the recommended course determination unit 122 may use the courses that have been optimized by the optimization calculations as recommended courses, or it may use only a portion of the courses that have been optimized by the optimization calculations as recommended courses.

[0029] As described above, the course support device 12 according to this exemplary embodiment employs a configuration that includes a data acquisition unit 121 that acquires state data indicating the status of the subject's course selection, and a recommended course determination unit 122 that determines the courses to recommend to the subject based on the results of an optimization calculation using an objective function generated by inverse reinforcement learning using training data indicating selected courses according to the status of course selection, and the state data acquired by the data acquisition unit 121.

[0030] According to the above configuration, state data indicating the subject's course selection status is used to determine recommended courses. This state data is then used in an optimization calculation using an objective function generated by inverse reinforcement learning with training data indicating selected courses corresponding to the subject selection status. Based on the results of this optimization calculation, recommended courses are determined. Therefore, the above configuration has the effect of making it possible to determine courses to recommend to a subject, taking into account their course selection status.

[0031] (Program for generating the objective function) The functions of the learning device 11 described above can also be implemented by a program. The objective function generation program according to this exemplary embodiment causes the computer to function as a data acquisition means for acquiring training data indicating the state of course selection and the courses selected according to that state, and as a learning means for generating an objective function for determining courses to recommend to a subject, according to the state of the subject selection, by performing inverse reinforcement learning using the acquired training data. This objective function generation program has the effect of making it possible to determine courses to recommend to a subject, taking into account the state of the subject selection.

[0032] (Course registration support program) Similarly, the functions of the course support device 12 described above can also be implemented by a program. In this exemplary embodiment, the course support program uses a computer as a data acquisition means to acquire state data indicating the state of a subject's course selection, and a recommended course determination means to determine which courses to recommend to the subject based on the results of an optimization calculation using an objective function generated by inverse reinforcement learning using training data indicating the state of course selection and the selected courses corresponding to that state, and the state data acquired by the data acquisition means. This course support program has the effect of making it possible to determine which courses to recommend to the subject, taking into account the state of the subject's course selection.

[0033] (The process for generating the objective function) The flow of the objective function generation method according to this exemplary embodiment will be explained with reference to Figure 2. Figure 2 is a flowchart showing the flow of the objective function generation method. Note that the entity executing each step in this objective function generation method may be a processor provided in the learning device 11, a processor provided in another device, or the entity executing each step may be a processor provided in a different device.

[0034] In S111 (data acquisition process), at least one processor acquires training data indicating the status of course selection and the selected courses corresponding to that status.

[0035] In S112 (learning process), at least one processor uses the training data acquired in S111 to perform inverse reinforcement learning, thereby generating an objective function for determining which subjects to recommend to the subject, based on the subject's state regarding their subject selection.

[0036] As described above, the objective function generation method according to this exemplary embodiment employs a configuration in which at least one processor performs a data acquisition process to acquire training data indicating the state of course selection and the courses selected according to that state, and a learning process to generate an objective function for determining courses to recommend to the subject, according to the state of the subject selection, by performing inverse reinforcement learning using the training data acquired in the data acquisition process. This objective function generation method has the effect of making it possible to determine courses to recommend to the subject, taking into account the state of the subject selection.

[0037] (Flow of course registration support methods) The flow of the course support method according to this exemplary embodiment will be explained with reference to Figure 3. Figure 3 is a flowchart showing the flow of the course support method. Note that the entity executing each step in this course support method may be a processor provided in the course support device 12, a processor provided in another device, or the entity executing each step may be a processor provided in a different device.

[0038] In S121 (data acquisition process), at least one processor acquires status data indicating the status of the subject's course selection.

[0039] In S122 (Recommended Course Determination Process), at least one processor determines which courses to recommend to the subject based on the results of an optimization calculation using an objective function generated by inverse reinforcement learning using training data that shows the state regarding course selection and the courses selected according to that state, and the state data obtained in S121.

[0040] As described above, the course support method according to this exemplary embodiment employs a configuration in which at least one processor performs a data acquisition process to acquire state data indicating the state of the subject's course selection, and a recommended course determination process to determine which courses to recommend the subject to take based on the results of an optimization calculation using an objective function generated by inverse reinforcement learning using training data indicating the state of course selection and the selected courses corresponding to that state, and the state data acquired in the data acquisition process. This course support method has the effect of making it possible to determine which courses to recommend the subject to take, taking into account the state of the subject's course selection.

[0041] [Second exemplary embodiment] A second exemplary embodiment, which is an example of an embodiment of the present invention, will be described in detail with reference to the drawings. Components having the same function as those described in the above-described exemplary embodiment are denoted by the same reference numerals, and their descriptions are omitted as appropriate. The scope of application of each technology adopted in this exemplary embodiment is not limited to this exemplary embodiment. That is, each technology adopted in this exemplary embodiment can also be adopted in other exemplary embodiments included in this disclosure, to the extent that no particular technical problems arise. Furthermore, each technology shown in the drawings referenced to describe this exemplary embodiment can also be adopted in other exemplary embodiments included in this disclosure, to the extent that no particular technical problems arise.

[0042] (Configuration of Course Registration Support System 2) The configuration of the course registration support device 2 according to this exemplary embodiment will be explained with reference to Figure 4. Figure 4 is a block diagram showing the configuration of the course registration support device 2. The course registration support device 2 is a device equipped with a function to assist a student in selecting courses. The course registration support device 2 also has a function to generate an objective function used to assist a student in selecting courses.

[0043] As shown in the figure, the course support device 2 comprises a control unit 20 that centrally controls all parts of the course support device 2, and a storage unit 21 that stores various data used by the course support device 2. The course support device 2 also comprises a communication unit 22 for the course support device 2 to communicate with other devices, an input unit 23 that receives input to the course support device 2, and an output unit 24 for the course support device 2 to output data. The control unit 20 includes a data acquisition unit 201, a reception unit 202, an objective function selection unit 203, an optimization calculation unit 204, a recommended course determination unit 205, a presentation control unit 206, and a learning unit 207.

[0044] The data acquisition unit 201 acquires various data necessary to support the subject's selection of courses. For example, the data acquisition unit 201 acquires state data indicating the subject's status regarding course selection, similar to the data acquisition unit 121 in the exemplary embodiment 1. Also, for example, the data acquisition unit 201 acquires training data indicating the status regarding course selection and the courses selected according to that status, similar to the data acquisition unit 111 in the exemplary embodiment 1.

[0045] The reception unit 202 accepts various inputs for course selection support. For example, the reception unit 202 accepts adjustments to the weight values ​​of the objective function used to determine recommended courses. Also, for example, the reception unit 202 accepts inputs regarding the subject's course selection requests.

[0046] The objective function selection unit 203 selects an objective function appropriate for the target individual from among multiple objective functions generated using different training data. Note that the objective function selection unit 203 is omitted when only one objective function is used.

[0047] The method for selecting the objective function is not particularly limited, as long as it allows for the selection of an objective function appropriate to the target user. For example, the objective function selection unit 203 may select an objective function from among several objective functions, specified by the user of the learning support device 2 (who may be the target user of the learning support or another person), as the objective function appropriate to the target user. Alternatively, for example, if an objective function is provided for each target user, the objective function selection unit 203 may identify the target user by obtaining their identification information and select the objective function for the identified target user.

[0048] The optimization calculation unit 204 performs an optimization calculation using an objective function generated by inverse reinforcement learning using training data that shows the state of course selection and the selected courses corresponding to that state, and state data acquired by the data acquisition unit 201. Details of the optimization calculation will be described later.

[0049] The recommended course determination unit 205, similar to the recommended course determination unit 122 in Exemplary Embodiment 1, determines the courses that the target user should take (recommended courses) based on the results of an optimization calculation using an objective function generated by inverse reinforcement learning using training data that shows the state regarding course selection and the selected courses corresponding to that state, and state data acquired by the data acquisition unit 201. This optimization calculation is performed by the optimization calculation unit 204. Details of the method for determining recommended courses will be described later.

[0050] The display control unit 206 displays various information related to supporting the selection of courses. For example, the display control unit 206 displays the recommended courses determined by the recommended course determination unit 205. The target audience for the display and the method of display are arbitrary. For example, the display control unit 206 may display the information to the user of the course support device 2 by outputting the information to the output unit 24. Alternatively, for example, the display control unit 206 may display the information to the target audience by outputting the information to the terminal device used by the target audience via communication through the communication unit 22. Furthermore, for example, if the target audience is a student, the display control unit 206 may display the information via an on-campus system available to that student.

[0051] The learning unit 207, similar to the learning unit 112 in Exemplary Embodiment 1, generates an objective function for determining recommended courses based on the subject's course selection status by performing inverse reinforcement learning using training data. The generated objective function is used in the optimization calculation by the optimization calculation unit 204. Details of the method for generating the objective function will be described later.

[0052] As described above, the course support device 2 according to this exemplary embodiment, like the course support device 12 of exemplary embodiment 1, includes a data acquisition unit 201 that acquires state data indicating the state of the subject's course selection, and a recommended course determination unit 205 that determines the courses to recommend to the subject based on the results of an optimization calculation using an objective function generated by inverse reinforcement learning using training data indicating selected courses according to the state of the course selection, and the state data acquired by the data acquisition unit 201. Therefore, the course support device 2 has the effect of being able to determine the courses to recommend to the subject, taking into account the state of the subject's course selection.

[0053] Furthermore, as described above, the course registration support device 2 is equipped with a reception unit 202 that accepts adjustments to the weight values ​​of the objective function. The weight values ​​of the objective function are values ​​multiplied by the explanatory variables corresponding to each perspective used to evaluate the appropriateness of the selection of courses. When an adjustment of the weight values ​​is accepted, the recommended course determination unit 205 determines the courses to recommend to the user based on the results of an optimization calculation performed using the adjusted weight values. Therefore, in addition to the effects of the course registration support device 12, the course registration support device 2 provides the effect of allowing users to adjust the perspectives that are emphasized and those that are not emphasized in determining recommended courses according to their preferences.

[0054] Furthermore, as described above, the course support device 2 includes an objective function selection unit 203 that selects an objective function appropriate to the subject from among multiple objective functions generated using different training data. The recommended course determination unit 205 then determines recommended courses based on the results of an optimization calculation using the objective function selected by the objective function selection unit 203. Therefore, in addition to the effects of the course support device 12, the course support device 2 provides the effect of being able to determine recommended courses that are more suitable for the subject.

[0055] Furthermore, as described above, the course support device 2 according to this exemplary embodiment, like the learning device 11 of exemplary embodiment 11, includes a data acquisition unit 201 that acquires training data indicating the state of course selection and the courses selected according to that state, and a learning unit 207 that generates an objective function for determining courses to be recommended to the subject, according to the state of the subject selection, by performing inverse reinforcement learning using the training data acquired by the data acquisition unit 201. Therefore, the course support device 2 has the effect of being able to determine courses to be recommended to the subject, taking into account the state of the subject selection. Note that it is not necessarily required that the generation of the objective function and the determination of recommended courses be performed by a single device.

[0056] (Method for generating the objective function) The method for generating the objective function by the learning unit 207 will be explained based on Figure 5. Figure 5 shows an example of objective function generation. The training data shown in Figure 5 includes subject data and student data as state data, as well as subject data indicating the subjects taken by the students.

[0057] Course data refers to data on each course that was a candidate for selection by the student, in other words, data on each course that was available to the student at the time of selection. The course data shown in Figure 5 includes the course name, the number of credits earned by taking the course, and the student's evaluation of the course. In addition to this, the course data may also include, for example, the percentage of students who obtained credits for each course, the grades of students taking each course, what other courses students took, the qualifications held by students taking each course, and the career paths of students taking each course. Career paths may indicate the type of occupation taken after graduation, such as researcher, salesperson, doctor, or medical professional, or they may indicate the company where the student was employed.

[0058] Student data refers to data about students enrolled in a course. The student data shown in Figure 5 includes the student's academic year at the time of enrollment, the subjects the student has already taken, and the student's grades in those subjects. Student data can be anything that indicates the characteristics of the student (which can also be rephrased as attributes, features, characteristics, etc.). For example, in addition to the data shown in Figure 5, student data could also include, for example, the student's strong subjects, weak subjects, career path, and qualifications held.

[0059] The above training data may be generated from the cases of students who actually selected and took courses and achieved excellent grades. More specifically, training data may be used that associates student data of high-achieving students with course data of courses that were available to those students when they selected their courses, and course data indicating the courses that those students took.

[0060] Learning Unit 207 can generate an objective function that reflects the criteria that high-achieving students prioritized when selecting courses, by using training data generated from the cases of high-achieving students. By using such an objective function, it becomes possible to determine courses that are likely to yield good results as recommended courses.

[0061] Similarly, by using training data generated from the experiences of students who were satisfied with their course selection, it becomes possible to determine recommended courses that are expected to yield high satisfaction. Furthermore, by using training data generated from the experiences of students who pursued a specific career path, it becomes possible to determine recommended courses that are considered useful for pursuing that career path.

[0062] Furthermore, if a subject whose recommended courses are to be determined has previously made course selections, the learning unit 207 may generate an objective function for that subject using training data related to those selections. More specifically, the learning unit 207 may generate an objective function using training data that shows the state of the subject's past course selections and the courses selected according to that state. In this way, the learning unit 207 may generate an objective function tailored to each subject.

[0063] When the recommended course determination unit 205 determines recommended courses based on the results of an optimization calculation using the objective function described above, which is generated by inverse reinforcement learning using training data that shows the state of the subject's past course selections and the courses selected according to that state, in addition to the effects of the course support device 12, it becomes possible to determine recommended courses using the same criteria as when the subject makes their own choices.

[0064] For example, the learning unit 207 may generate an objective function for a student using training data that shows the subjects a student took in the first semester of a given academic year, the grades for those subjects, and the student's evaluation of those subjects. This allows the recommended subject determination unit 205 to determine, for example, the recommended subjects for the student's second semester selection, based on the results of an optimization calculation using the generated objective function, using the same criteria as when the student determined the recommended subjects for the first semester.

[0065] The training data only needs to indicate the subjects that should be selected in a given state, based on the student's current course selection status. For example, typical state data could be created, along with appropriate courses corresponding to each state, and these could be linked together to form the training data. Furthermore, constraints may be included in the training data as needed, regardless of whether the training data is based on the actual course selection results of students.

[0066] The training data described above is acquired by the data acquisition unit 201. Then, the learning unit 207 generates an objective function corresponding to the acquired training data. For example, the data acquisition unit 201 may acquire training data that includes the evaluation results of the subject by students who have taken the subject. This has the effect of generating an objective function that can determine recommended subjects to take, taking into account not only the effects of the learning device 11, but also the quality of the evaluation results by students who have actually taken the subject.

[0067] Furthermore, the data acquisition unit 201 may select training data to be used for inverse reinforcement learning from training data that shows the status of a student's course selection and the courses they have taken, based on the grades of multiple students who have taken the course. This provides the effect of generating an objective function that can determine recommended courses to take, taking into account grades, in addition to the effects of the learning device 11.

[0068] For example, the data acquisition unit 201 may acquire training data showing the status of course selection and the courses taken by a predetermined number of students who have a high average grade in the courses they have taken, among a group of students. This makes it possible to generate an objective function that learns the intentions of high-achieving students when they select courses. The same effect can be expected if the target students are, for example, students whose average grade in the courses they have taken is above a predetermined threshold.

[0069] The learning unit 207 performs inverse reinforcement learning using multiple training sets, each with different state data, to generate an objective function for determining which courses to recommend to the subject, based on the subject's state regarding course selection. For example, the learning unit 207 may generate an objective function that includes weight values ​​indicating how much importance to give to multiple perspectives for evaluating the validity of course selection, as shown in Figure 5. This learning can also be described as learning the intention behind a student's selection of a course data set when they are in a state represented by the state data included in that training set.

[0070] Here, "perspective" specifically corresponds to the explanatory variables in the objective function. The perspective (explanatory variable) can be anything related to the selection of courses, and the choice of which perspective to apply is arbitrary. For example, one might apply perspectives such as emphasizing credits, emphasizing desired career path, emphasizing grades, or emphasizing survey results (evaluations by students). Specifically, applying a perspective means applying explanatory variables to evaluate the selection of courses from that perspective.

[0071] For example, the objective function shown in Figure 5 includes "number of credits" as perspective 1, "desired career path" as perspective 2, and "grades" as perspective 3. Furthermore, each of these perspectives is multiplied by weights α, β, and γ, respectively. This means that the sum of the values ​​obtained by multiplying the explanatory variable for evaluation from the perspective of "number of credits" by weight α, the explanatory variable for evaluation from the perspective of "desired career path" by weight β, and the explanatory variable for evaluation from the perspective of "grades" by weight γ is the evaluation value determined by the objective function. For example, the explanatory variable for evaluation from the perspective of "number of credits" could be the total number of credits obtained from the selected set of courses. For example, the explanatory variable for evaluation from the perspective of "desired career path" could be the degree of fit between the desired career path and the actual career path. For example, the explanatory variable for evaluation from the perspective of "grades" could be the average grade for the selected set of courses.

[0072] In the learning process, first, the learning unit 207 sets the weight values ​​multiplied by each explanatory variable of the objective function to their initial values. Next, the optimization calculation unit 204 performs an optimization calculation using the state data included in the training data and the objective function with the weight values ​​set to their initial values. Subsequently, the recommended course determination unit 205 determines the recommended courses to take based on the results of the optimization calculation. Then, the learning unit 207 updates the weight values ​​so that the difference between the courses shown in the training data and the recommended courses determined by the recommended course determination unit 205 becomes small. By repeating these processes until the difference between the courses shown in the training data and the recommended courses determined by the recommended course determination unit 205 becomes sufficiently small, the learning of the objective function is completed.

[0073] As for specific learning methods, various techniques commonly used in inverse reinforcement learning can be applied. For example, the maximum entropy inverse reinforcement learning method can be applied. In this case, the learning unit 207 uses the maximum entropy principle to represent the probability distribution of the objective function and learns the objective function by bringing the probability distribution of the objective function closer to the true probability distribution (i.e., maximum likelihood estimation). Furthermore, it is possible to identify appropriate perspectives for evaluating the courses included in the training data through learning.

[0074] (Generating training data) The data acquisition unit 201 may acquire training data generated outside the learning support device 2 via the communication unit 22 or the input unit 23, or it may generate training data and acquire the generated training data. When generating training data, the data acquisition unit 201 first acquires the state data and course data necessary for generating the training data.

[0075] For example, when acquiring state data included in the training data in Figure 5, the data acquisition unit 201 may acquire subject data such as subject name and number of credits from data indicating the curriculum, such as a syllabus (in digital form). Alternatively, if such subject data is recorded in the academic affairs system of the educational institution to which the student belongs, the data acquisition unit 201 may acquire such subject data from the academic affairs system. Furthermore, the data acquisition unit 201 can also acquire student data and course data from various systems belonging to the educational institution, such as the academic affairs system or a learning management system that manages students' learning progress. Additionally, student evaluations of subjects may be obtained through surveys targeting students, in which case the data acquisition unit 201 can acquire student evaluations of subjects from a database recording the survey results.

[0076] Furthermore, the student enrollment data may contain personal information. For this reason, the data acquisition unit 201 may acquire student enrollment data after anonymizing the personal information (such as students' names) in the data recorded in the system described above. Examples of anonymization processes include removing personal information, masking personal information, and encrypting personal information. The data acquisition unit 201 may also perform the anonymization process itself.

[0077] Having acquired the state data and course data as described above, the data acquisition unit 201 generates training data by associating the acquired state data and course data. If the learning unit 207 generates multiple objective functions, the data acquisition unit 201 generates training data corresponding to each objective function.

[0078] (Method for determining recommended courses) Using an objective function, an evaluation value can be calculated to assess the quality of the course selection. Therefore, the optimization calculation performed by the optimization calculation unit 204 is to solve an optimization problem that finds the combination of courses that maximizes the evaluation value calculated using the objective function. Then, the recommended course determination unit 205 simply determines the combination of courses found by the optimization calculation unit 204 as the recommended courses. The method for solving the optimization problem using the objective function, state data, and constraints is arbitrary. For example, the optimization calculation unit 204 may use an optimization solver to find the optimal combination of courses from the objective function, state data, and constraints.

[0079] Figure 6 shows an example of determining recommended courses. In the example in Figure 6, the input data used to determine recommended courses includes course data showing multiple courses that the student can choose from and the content of those courses, student data showing the characteristics of the student, and constraints on course selection. Of these, the course data and student data are state data that shows the student's status regarding course selection.

[0080] Status data may be entered by the user of the learning support device 2, such as the target student. Alternatively, the data acquisition unit 201 may automatically acquire some or all of the status data. In this case, for example, the user may input their own identification information into the learning support device 2, and the data acquisition unit 201 may acquire the target student data and subject data of the target student identified by that identification information from a predetermined database. Furthermore, the input of constraint conditions can be received by the reception unit 202.

[0081] Examples of the above-mentioned databases include academic affairs system databases, databases recording student survey results, databases recording course evaluation results (e.g., questionnaire results), databases recording syllabus data, databases for systems managing student enrollment and grades, and databases for other systems owned by educational institutions.

[0082] The subject data shown in Figure 6 includes the names of the subjects available to the target students, the number of credits earned by taking those subjects, and the students' evaluations of those subjects. In addition, the subject data may also include, for example, the percentage of students who earned credits for each subject, the grades of students taking each subject, what other subjects students have taken, the qualifications held by students taking each subject, and the career paths of students taking each subject.

[0083] The data acquisition unit 201 can acquire data such as course name and number of credits from, for example, a syllabus (in digital format). Furthermore, the data acquisition unit 201 can acquire student evaluations of courses from a database containing the results of surveys conducted among students.

[0084] Furthermore, the subject data shown in Figure 6 includes the subject's academic year, completed subjects, and grades for those subjects. Such data can be obtained, for example, from a database of a system that manages students' course registration status and grades. The subject data should represent the subject's characteristics and should include the same content as the course registration data. For example, in addition to the data shown in Figure 6, subject data could also include the subject's strong subjects, weak subjects, desired career path, and qualifications the subject wishes to obtain.

[0085] Furthermore, the constraints shown in Figure 6 include the requirement that students must select all required courses they have not yet taken, and that Saturdays should be days off. Constraints can be set arbitrarily. Constraints can also be set in advance, or students can set them themselves. Students can also change the constraints.

[0086] For example, the reception unit 202 may accept input regarding the subject's course selection. The recommended course determination unit 205 may then determine recommended courses for the subject based on the results of an optimization calculation using the received requests as constraints. This provides the added benefit of determining recommended courses that reflect the subject's preferences, in addition to the effects of the course support device 12.

[0087] The optimization calculation unit 204 finds the combination of subjects shown in the subject data in the state data described above that satisfies the constraints and yields the best evaluation using the objective function. Then, the recommended subject determination unit 205 determines the recommended subjects to be taken based on the results of the optimization calculation. For example, the recommended subject determination unit 205 may use each subject included in the combination deemed optimal by the optimization calculation as the recommended subjects to be taken, or it may use only some of the subjects included in the combination deemed optimal by the optimization calculation as the recommended subjects to be taken.

[0088] When selecting recommended subjects from a combination determined to be optimal through optimization calculations, the criteria for selecting these subjects can be predetermined. For example, the recommended subject determination unit 205 may use a list of subjects recommended by educational institutions and select the subjects recommended by educational institutions from among the subjects included in the combination determined to be optimal through optimization calculations.

[0089] The recommended courses determined by the recommended course determination unit 205 are presented to the target user by the presentation control unit 206. The method of presenting the recommended courses is arbitrary. For example, the presentation control unit 206 may display the recommended courses for each day of the week on the display device, as shown in the example in Figure 6. Alternatively, the presentation control unit 206 may display the recommended courses in a format similar to a timetable. In this case, the presentation control unit 206 should display the recommended courses for each time slot on each day of the week in a timetable format divided by day of the week and time slot.

[0090] (Process flow) The processes (course support methods) executed by the course support device 2 will be explained based on Figure 7. Figure 7 is a flowchart showing the flow of processes executed by the course support device 2. The processes shown in Figure 7 include each process of the course support method according to this exemplary embodiment.

[0091] In S21, the reception unit 202 receives the identification information of the subject and the subject's requests regarding course selection. The entered requests are used as constraints in the optimization calculation, as described above. Furthermore, the entered requests can also be used in selecting the objective function.

[0092] In S22 (data acquisition process), the data acquisition unit 201 acquires status data (data indicating the status related to course selection) of the subject identified by the identification information received in S21. As mentioned above, status data can be acquired from various databases. Alternatively, the user may be required to input both status data and requests.

[0093] In S23, the objective function selection unit 203 selects an objective function appropriate to the subject from among multiple objective functions, each generated using different training data. For example, if an objective function is prepared for each subject, the objective function selection unit 203 selects the objective function for the subject identified by the identification information received in S21. Alternatively, if an objective function is prepared for each subject's request, the objective function selection unit 203 selects the objective function for the subject who has the request received in S21. For example, for a subject who has a request to pursue a specific career path, the objective function selection unit 203 may select an objective function generated using training data of students who have pursued that career path.

[0094] In S24, the presentation control unit 206 presents the objective function selected in S23 to the user. At this time, it is preferable that the presentation control unit 206 expresses the perspectives in the objective function using text or the like, as shown in Figure 5, and also explicitly indicates weight values, which are the values ​​to be adjusted, to show the degree to which each perspective for evaluating the appropriateness of the selection of courses is given importance.

[0095] In S25, the reception unit 202 accepts adjustments to the weight values ​​multiplied by the explanatory variables corresponding to each viewpoint in the objective function. If there is no need to adjust the weight values, the processes in S24 and S25 are omitted. Alternatively, after S24, the reception unit 202 may accept an operation to change the objective function presented in S24 to a different objective function.

[0096] In S26, the optimization calculation unit 204 performs an optimization calculation using an objective function generated by inverse reinforcement learning using training data that shows the state of course selection and the courses selected according to that state, and the state data obtained in S22. Then, in S27 (recommended course determination process), the recommended course determination unit 205 determines the courses that the subject should take based on the results of the optimization calculation performed in S26.

[0097] In S28, the presentation control unit 206 presents the recommended courses determined in S27. This completes the process shown in Figure 7. The reception unit 202 may accept adjustments to weight values ​​and additional input of requests after S28. The reception unit 202 may also accept changes or re-inputs of at least one of the state data and / or requests after S28. If the data used in the optimization calculation is changed in this way, the processing from S26 onwards is performed based on the changed data.

[0098] Furthermore, the learning support device 2 is equipped with a learning unit 207. Therefore, if a training dataset tailored to a target individual can be obtained or generated at the time the target individual is identified, it is possible to have the learning unit 207 generate an objective function tailored to that target individual at that time. In this case, the optimization calculation in S26 uses the objective function tailored to the target individual, which is generated by the learning unit 207.

[0099] [Variation] The entities executing each process described in the exemplary embodiments above are arbitrary and not limited to the examples given. For example, a learning support system having the same functions as learning support device 2 can be constructed using multiple devices that can communicate with each other. For example, a learning support system having the same functions as learning support device 2 can be constructed by distributing each block shown in Figure 4 across multiple devices. Furthermore, the entities executing each process shown in the flowchart of Figure 7 may be a single device (which can also be called a processor) or multiple devices (which can also be called processors).

[0100] [Examples of implementation using software] Some or all of the functions of the learning device 11, the course support device 12, and the course support device 2 (hereinafter also referred to as "each of the above devices") may be implemented by hardware such as integrated circuits (IC chips) or by software.

[0101] In the latter case, each of the above devices is implemented, for example, by a computer that executes instructions for a program, which is software that realizes each function. An example of such a computer (hereinafter referred to as Computer C) is shown in Figure 8. Figure 8 is a block diagram showing the hardware configuration of Computer C, which functions as each of the above devices.

[0102] Computer C comprises at least one processor C1 and at least one memory C2. Memory C2 stores a program (study support program / objective function generation program) P that causes computer C to operate as each of the above-mentioned devices. In computer C, processor C1 reads program P from memory C2 and executes it, thereby realizing each of the above-mentioned functions.

[0103] For processor C1, for example, a CPU (Central Processing Unit), GPU (Graphic Processing Unit), DSP (Digital Signal Processor), MPU (Micro Processing Unit), FPU (Floating Point Number Processing Unit), PPU (Physics Processing Unit), TPU (Tensor Processing Unit), quantum processor, microcontroller, or a combination thereof can be used. For memory C2, for example, flash memory, HDD (Hard Disk Drive), SSD (Solid State Drive), or a combination thereof can be used.

[0104] Computer C may also be equipped with RAM (Random Access Memory) for loading program P at runtime and for temporarily storing various data. Furthermore, computer C may be equipped with communication interfaces for sending and receiving data with other devices. Additionally, computer C may be equipped with input / output interfaces for connecting input / output devices such as keyboards, mice, displays, and printers.

[0105] Furthermore, program P can be recorded on a non-temporary, tangible recording medium M that is readable by computer C. Such a recording medium M could be, for example, tape, disk, card, semiconductor memory, or programmable logic circuitry. Computer C can acquire program P via such a recording medium M. Program P can also be transmitted via a transmission medium. Such a transmission medium could be, for example, a communication network or broadcast waves. Computer C can also acquire program P via such a transmission medium.

[0106] Furthermore, each of the above functions of each of the above devices may be implemented by a single processor in a single computer, by multiple processors in a single computer working together, or by multiple processors in each of multiple computers working together. In addition, the programs for implementing each of the above functions in each of the above devices may be stored in a single memory in a single computer, distributed and stored in multiple memories in a single computer, or distributed and stored in multiple memories in each of multiple computers.

[0107] [Additional Notes] This disclosure includes the technologies described in the following appendices. However, the present invention is not limited to the technologies described in the following appendices, and various modifications are possible within the scope of the claims.

[0108] (Note A1) A course enrollment support device comprising: a data acquisition means for acquiring state data indicating the status of a subject's course selection; a recommended course determination means for determining courses to recommend to the subject based on the results of an optimization calculation using an objective function generated by inverse reinforcement learning using training data indicating the status of course selection and the selected courses corresponding to that status, and the state data acquired by the data acquisition means.

[0109] (Appendix A2) The course support device described in Appendix A1, comprising a reception means for accepting adjustments to weight values ​​multiplied by explanatory variables corresponding to each perspective for evaluating the appropriateness of the selection of courses in the objective function, wherein the recommended course determination means determines the courses to be recommended to the subject based on the results of an optimization calculation performed by applying the weight values ​​after the adjustments have been made.

[0110] (Note A3) The learning support device described in Appendix A1 or A2, comprising: an objective function selection means for selecting an objective function appropriate to the subject from among multiple objective functions each generated using different training data; and a recommended subject determination means for determining subjects to recommend to the subject based on the results of an optimization calculation using the objective function selected by the objective function selection means.

[0111] (Note A4) A course support device as described in any of Appendix A1 to A3, comprising a receiving means for receiving input of the subject's requests regarding course selection, and a recommended course determination means for determining courses to recommend to the subject based on the results of an optimization calculation with the requests as constraints.

[0112] (Note A5) The recommended subject determination means determines the subjects to be recommended for the subject based on the results of an optimization calculation using an objective function generated by inverse reinforcement learning using training data that shows the state of the subject's past subject selections and the subjects selected according to that state, and state data acquired by the data acquisition means, as described in any of Appendix A1 to A4.

[0113] (Note A6) A learning device comprising: data acquisition means for acquiring training data indicating the state of course selection and the courses selected according to that state; and learning means for generating an objective function for determining courses to recommend to a subject according to the subject's state of course selection by performing inverse reinforcement learning using the training data.

[0114] (Note A7) The data acquisition means is a learning device as described in Appendix A6, which acquires training data including evaluation results of the subject by students who have taken the subject.

[0115] (Note A8) The data acquisition means is a learning device as described in Appendix A6 or A7, which selects training data to be used for inverse reinforcement learning from training data that shows the status of a student's course selection and the courses they have taken, based on the grades of multiple students who have taken the course.

[0116] (Note B1) A course enrollment support method comprising: a data acquisition process in which at least one processor acquires state data indicating the state of a subject's course selection; and a recommended course determination process that determines which courses to recommend the subject to take based on the results of an optimization calculation using an objective function generated by inverse reinforcement learning using training data indicating the state of course selection and the courses selected according to that state, and the state data acquired in the data acquisition process.

[0117] (Note B2) The course enrollment support method according to Appendix B1, wherein the at least one processor includes an acceptance process for accepting adjustments to weight values ​​multiplied by explanatory variables corresponding to each perspective for evaluating the appropriateness of the selection of courses in the objective function, and in the recommended course determination process, the at least one processor determines the courses to be recommended to the subject based on the results of an optimization calculation performed by applying the weight values ​​after the adjustments have been made.

[0118] (Note B3) The course support method according to Appendix B1 or B2, wherein the at least one processor includes an objective function selection process in which it selects an objective function appropriate to the subject from among a plurality of objective functions each generated using different training data, and in the recommended course determination process, the at least one processor determines the subjects to be recommended for the subject to take based on the result of an optimization calculation using the objective function selected in the objective function selection process.

[0119] (Note B4) The course support method according to any one of Appendix B1 to B3, wherein the at least one processor includes a reception process for receiving input of a subject's request regarding the selection of courses to take, and in the recommended course determination process, the at least one processor determines a course to recommend to the subject based on the result of an optimization calculation with the request as a constraint.

[0120] (Note B5) In the recommended course determination process, the course support method described in any of Appendix B1 to B4, wherein the at least one processor determines the courses to recommend to the subject based on the result of an optimization calculation using an objective function generated by inverse reinforcement learning using training data that shows the state of the subject's past course selections and the courses selected according to that state, and the state data obtained in the data acquisition process.

[0121] (Note B6) A method for generating an objective function, comprising: a data acquisition process in which at least one processor acquires training data indicating a state regarding course selection and the courses selected according to that state; and a learning process that generates an objective function for determining courses to recommend to a subject according to the state regarding course selection by performing inverse reinforcement learning using the training data.

[0122] (Note B7) The method for generating the objective function described in Appendix B6, wherein the data acquisition process involves at least one processor acquiring training data including evaluation results of the subject by students who have taken the subject.

[0123] (Note B8) In the data acquisition process, the method for generating the objective function described in Appendix B6 or B7, wherein the at least one processor selects training data to be used for inverse reinforcement learning from training data that shows the status of a student's course selection and the courses they have taken, based on the grades of multiple students who have taken the course.

[0124] (Note C1) A program that causes a computer to function as a course registration support device, wherein the computer functions as a data acquisition means for acquiring state data indicating the state of a subject's course selection, and a recommended course determination means for determining courses to recommend to the subject based on the results of an optimization calculation using an objective function generated by inverse reinforcement learning using training data indicating the state of course selection and the selected courses corresponding to that state, and the state data acquired by the data acquisition means.

[0125] (Note C2) The course support program as described in Appendix C1, wherein the computer functions as a receiving means for receiving adjustments to the weight values ​​multiplied by the explanatory variables corresponding to each perspective for evaluating the appropriateness of the selection of courses in the objective function, and the recommended course determination means determines the courses to be recommended to the subject based on the results of an optimization calculation performed by applying the weight values ​​after the adjustments have been made.

[0126] (Note C3) The aforementioned computer functions as an objective function selection means for selecting an objective function appropriate to the subject from among a plurality of objective functions generated using different training data, and the recommended subject determination means determines the subjects to be recommended for the subject based on the results of an optimization calculation using the objective function selected by the objective function selection means, as described in Appendix C1 or C2 of the study support program.

[0127] (Note C4) A course support program as described in any of Appendix C1 to C3, wherein the computer functions as a means for receiving input of the subject's requests regarding the selection of courses to be taken, and the recommended course determination means determines the courses to be taken by the subject based on the results of an optimization calculation with the requests as constraints.

[0128] (Note C5) The aforementioned recommended course determination means is a course support program described in any of Appendix C1 to C4, which determines the courses to recommend to the subject based on the results of an optimization calculation using an objective function generated by inverse reinforcement learning using training data that shows the state of the subject's past course selections and the courses selected according to that state, and state data acquired by the data acquisition means.

[0129] (Appendix C6) A program that causes a computer to function as a learning device, wherein the computer functions as a data acquisition means for acquiring training data indicating the state of course selection and the courses selected according to that state, and a learning means for generating an objective function for determining courses to recommend to a subject according to the state of course selection of the subject, by performing inverse reinforcement learning using the training data.

[0130] (Note C7) The data acquisition means is a program for generating the objective function described in Appendix C6, which acquires training data including evaluation results of the subject by students who have taken the subject.

[0131] (Note C8) The data acquisition means is a program for generating the objective function described in Appendix C6 or C7, which selects training data to be used for inverse reinforcement learning from training data that shows the status of a student's course selection and the courses they have taken, based on the grades of multiple students who have taken the course.

[0132] (Note D1) A study support device comprising at least one processor, wherein the at least one processor performs a data acquisition process to acquire state data indicating the state of a subject's course selection, and a recommended course determination process to determine which courses to recommend the subject to take based on the results of an optimization calculation using an objective function generated by inverse reinforcement learning using training data indicating the state of course selection and the selected courses corresponding to that state, and the state data acquired in the data acquisition process.

[0133] The learning support device may also include memory. Furthermore, the memory may store a program that causes at least one processor to execute each of the aforementioned processes.

[0134] (Note D2) The course enrollment support device according to Appendix D1, wherein the at least one processor performs an acceptance process to accept adjustments to the weight values ​​multiplied by each explanatory variable corresponding to each perspective for evaluating the appropriateness of the selection of courses in the objective function, and in the recommended course determination process, the at least one processor determines the courses to be recommended to the subject based on the results of an optimization calculation performed by applying the weight values ​​after the adjustments have been made.

[0135] (Note D3) The learning support device according to Appendix D1 or D2, wherein at least one processor performs an objective function selection process to select an objective function appropriate to the subject from among a plurality of objective functions each generated using different training data, and in the recommended subject determination process, at least one processor determines the subjects to be recommended for the subject to take based on the results of an optimization calculation using the objective function selected in the objective function selection process.

[0136] (Note D4) The course enrollment support device according to any one of the appendices D1 to D3, wherein at least one processor performs an input processing for receiving the subject's request regarding the selection of courses to be taken, and in the recommended course determination processing, the at least one processor determines the courses to be taken by the subject based on the result of an optimization calculation with the request as a constraint.

[0137] (Note D5) In the recommended subject determination process, the at least one processor determines the subjects to be recommended for the subject based on the result of an optimization calculation using an objective function generated by inverse reinforcement learning using training data that shows the state of the subject's past subject selections and the subjects selected according to that state, and state data acquired in the data acquisition process, as described in any of the appendices D1 to D4.

[0138] (Note D6) A learning device comprising at least one processor, wherein the at least one processor performs a data acquisition process to acquire training data indicating a state regarding course selection and the courses selected according to that state, and a learning process to generate an objective function for determining courses to be recommended to a subject according to the state regarding course selection by performing inverse reinforcement learning using the training data.

[0139] (Note D7) In the data acquisition process, the learning device according to Appendix D6, wherein the at least one processor acquires training data including evaluation results of the subject by students who have taken the subject.

[0140] (Note D8) The learning device according to Appendix D6 or D7, wherein in the data acquisition process, the at least one processor selects training data to be used for inverse reinforcement learning from training data indicating the status of a student's course selection and the courses they have taken, based on the grades of multiple students who have taken the course.

[0141] (Note E1) A non-temporary recording medium that records a program for causing a computer to function as a course registration support device, the program causing the computer to execute: a data acquisition process for acquiring state data indicating the state of a subject's course selection; a recommended course determination process for determining which courses to recommend the subject to take based on the results of an optimization calculation using an objective function generated by inverse reinforcement learning using training data indicating the state of course selection and the selected courses corresponding to that state, and the state data acquired in the data acquisition process.

[0142] (Note E2) A non-temporary recording medium that records a program for causing a computer to function as a learning device, the program for causing the computer to execute a data acquisition process that acquires training data indicating the state of the subject's course selection and the subject selected according to that state, and a learning process that generates an objective function for determining which subjects to recommend the subject to take according to the subject's state of course selection by performing inverse reinforcement learning using the training data. [Explanation of Symbols]

[0143] 11 Learning device 111 Data acquisition unit (data acquisition means) 112 Learning Section (Learning Methods) 12. Course registration support system 121 Data acquisition unit (data acquisition means) 122 Recommended Subject Determination Unit (Method for Determining Recommended Subjects) 2. Course registration support devices (learning devices) 201 Data acquisition unit (data acquisition means) 202 Reception area (reception method) 203 Objective function selection unit (objective function selection means) 207 Learning Section (Learning Methods)

Claims

1. A data acquisition means for acquiring status data that shows the status of the subject's course selection, A course registration support device comprising: an objective function generated by inverse reinforcement learning using training data indicating the state of course selection and the courses selected according to that state; and a recommended course determination means that determines courses to recommend to the subject based on the results of an optimization calculation using state data acquired by the data acquisition means.

2. The aforementioned objective function includes a means for accepting adjustments to the weight values ​​multiplied by the explanatory variables corresponding to each perspective for evaluating the appropriateness of the selection of courses, The course selection support device according to claim 1, wherein the recommended course selection means determines the courses to be recommended for the subject based on the results of an optimization calculation performed by applying the weight values ​​after the adjustments have been made.

3. The system includes an objective function selection means that selects an objective function appropriate to the subject from among multiple objective functions generated using different training data. The course selection support device according to claim 1 or 2, wherein the recommended course selection means determines the courses to be recommended for the subject based on the results of an optimization calculation using the objective function selected by the objective function selection means.

4. The system includes a means for receiving input regarding the subject selection of the aforementioned individuals, The course selection support device according to claim 1 or 2, wherein the recommended course selection means determines the courses to be recommended for the subject based on the results of an optimization calculation with the request as a constraint.

5. The recommended subject determination means determines subjects to be recommended for the subject to take based on the results of an optimization calculation using an objective function generated by inverse reinforcement learning using training data that shows the state of the subject's past subject selections and the subjects selected according to that state, and state data acquired by the data acquisition means, according to claim 1 or 2.

6. A data acquisition means for acquiring training data that shows the status regarding course selection and the selected courses corresponding to that status, A learning device comprising: a learning means for generating an objective function for determining which subjects to recommend a subject to take, according to the subject's state regarding subject selection, by performing inverse reinforcement learning using the aforementioned training data.

7. The learning device according to claim 6, wherein the data acquisition means acquires training data including evaluation results of the subject by students who have taken the subject.

8. The learning device according to claim 6 or 7, wherein the data acquisition means selects training data to be used for inverse reinforcement learning from training data indicating the status of a student's course selection and the courses they have taken, based on the grades of multiple students who have taken the course.

9. At least one processor, A data acquisition process that obtains status data indicating the status of the subject's course selection, A course enrollment support method that performs a recommended course determination process to determine which courses to recommend to the subject based on the results of an optimization calculation using an objective function generated by inverse reinforcement learning using training data that shows the state regarding course selection and the courses selected according to that state, and state data obtained in the data acquisition process.

10. At least one processor, A data acquisition process that obtains training data showing the status of course selection and the selected courses corresponding to that status, A method for generating an objective function, which involves performing a learning process to generate an objective function for determining which subjects to recommend a subject to take, according to the subject's state regarding subject selection, by performing inverse reinforcement learning using the aforementioned training data.

Images