Information processing device, support method, and support program

Abstract

To reduce the burden of manual work in life planning. [Solution] The information processing device comprises a classification unit that classifies multiple numerical data related to a subject's life plan into data items for generating life planning support information using content information and a language model that indicate the content of each numerical data, and a support information generation unit that generates support information using the numerical data as data for the data items classified by the classification unit. Furthermore, this information processing device can also support the subject's decision-making.

Description

Technical Field

[0001] The present disclosure relates to an information processing apparatus, a support method, and a support program.

Background Art

[0002] Recently, there has been an increasing interest in life planning. In life planning, a life plan including a fund plan from the present to the future is created based on the current assets and income and expenditure situation, family composition, and future design of the target person.

[0003] As a technology for supporting life planning, for example, the information processing system described in Patent Document 1 can be cited. In the information processing system described in Patent Document 1, future predictions are made for each of the health assets and economic assets of the target person, and information on options that the target person can select is generated based on the prediction results.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] The information processing system described in Patent Document 1 uses data obtained by quantifying the total amount of economic assets such as money, land, buildings, and securities belonging to the target person for future prediction of economic assets. However, it takes time and effort to prepare such data. Specifically, in order to create data obtained by quantifying the total amount of economic assets, the target person is forced to perform complicated operations such as checking the deposit and withdrawal details of their own accounts, classifying current income and expenditure into salary income, housing expenses, etc., and aggregating the amounts for each classification. Thus, the information processing system described in Patent Document 1 has room for improvement in that the burden of manual work in life planning is large.

[0006] This disclosure has been made in view of the above-mentioned issues, and one exemplary purpose is to provide a technology that can reduce the burden of manual work in life planning. [Means for solving the problem]

[0007] An information processing device relating to an illustrative aspect of this disclosure includes a classification means that classifies multiple numerical data relating to a subject's life plan into data items for generating support information to support the subject's life planning, using content information indicating the content of each numerical data and a language model that has been machine-learned from natural language; and a support information generation means that generates the support information using the numerical data as data for the data items classified by the classification means.

[0008] In an exemplary aspect of the present disclosure, at least one processor performs a classification process that classifies multiple numerical data relating to a subject's life plan into data items for generating support information to assist the subject's life planning, using content information indicating the content of each numerical data and a language model trained on natural language learning; and a support information generation process that generates the support information using the numerical data as data for the data items classified in the classification process.

[0009] An exemplary support program relating to this disclosure causes a computer to function as a classification means for classifying multiple numerical data related to a subject's life plan into data items for generating support information to assist the subject's life planning, using content information indicating the content of each numerical data and a language model trained on natural language. A support information generation means generates the support information by using the numerical data as data for the data items classified by the classification means. [Effects of the Invention]

[0010] One illustrative aspect of this disclosure is that it can reduce the burden of manual work in life planning. [Brief explanation of the drawing]

[0011] [Figure 1] This is a block diagram showing the configuration of the information processing device related to this disclosure. [Figure 2] This flowchart shows the flow of support methods related to this disclosure. [Figure 3] This figure shows an example of processing performed by other information processing devices related to this disclosure. [Figure 4] Figure 3 is a block diagram showing the configuration of the information processing device. [Figure 5] This figure shows a prompt generated by the classification unit and an example of the inference result obtained by inputting that prompt into the language model. [Figure 6] This diagram illustrates the presentation and reclassification of inference results. [Figure 7] Figure 3 is a flowchart showing the processing flow executed by the information processing device. [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] (Configuration of Information Processing Device 1) The configuration of the information processing device 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 information processing device 1. As shown in Figure 1, the information processing device 1 includes a classification unit 101 and a support information generation unit 102.

[0015] The classification unit 101 classifies a plurality of numerical data related to the life plan of the target person into each data item for generating support information for assisting the life planning of the target person, using content information indicating the content of each numerical data and a language model obtained by machine learning of natural language.

[0016] The plurality of numerical data related to the life plan may be any data used for generating the support information. Also, the content information may be any information indicating the content of each numerical data. For example, the deposit and withdrawal details (which can also be referred to as the deposit and withdrawal history) in the bank accounts and securities accounts held by the target person include numerical data such as the amount of deposits, the amount of withdrawals, and the account balance, as well as text data indicating the content. Therefore, the classification unit 101 can also classify the numerical data shown in the deposit and withdrawal details using the text data shown in the deposit and withdrawal details. In addition, the classification unit 101 can classify, for example, numerical data obtained from credit card statements, electronic money transaction histories, public utility payment statements, medical expense statements, numerical data extracted from receipts, personal identification cards, withholding tax certificates, tax returns, tax payment certificates, tax assessment certificates, and loan balance certificates, etc. Specific numerical data includes, for example, the amount of deposits, the amount of withdrawals, the account balance, the amount of expenditure, the amount of income, the amount of tax paid, the amount of income, the loan balance, the amount of assets held, and the amount of tax deductions, etc. Also, as numerical data other than the amount, for example, data indicating various activities of the target person such as sleep time, working time, exercise time, etc., and data indicating the health status of the target person such as weight, blood pressure, blood sugar level, etc. can be mentioned.

[0017] Also, the content information indicating the content of the numerical data may be displayed or recorded in association with the numerical data, or may be obtained separately from the numerical data. The content information may be, for example, text data indicating the content of the numerical data in natural language, or data such as an image indicating the content of the numerical data. When using data such as an image as the content information, the language model to be used may be a model that can take data such as an image as input data.

[0018] Here, learning a natural language more specifically means learning the arrangement of its components (such as words) in a sentence of the natural language and the arrangement of sentences in a text. Examples of language models that have learned a natural language include, for example, BERT (Bidirectional Encoder Representations from Transformers), RoBERTa (Robustly optimized BERT approach), ELECTRA (Efficiently Learning an Encoder that Classifies Token Replacements Accurately), and the like.

[0019] Note that the above language model may be a general-purpose language model that can be used for purposes other than the classification of numerical data, or it may be a general-purpose language model fine-tuned for the classification of numerical data. Also, the language model may be provided in the information processing apparatus 1 or in another apparatus. In the latter case, the classification unit 101 uses the language model via another apparatus equipped with the language model.

[0020] The support information generation unit 102 generates support information using the numerical data regarding the life plan of the target person as the data of the data items classified by the classification unit 101.

[0021] Here, the above support information can be generated using the data of a predetermined data item, and any information that can be used to support the life planning of the target person is acceptable. For example, the above support information may be information indicating the simulation results of the future income and expenditure of the target person. Such a simulation can be performed using the current income amount and expenditure amount of the target person and the like.

[0022] Furthermore, the above-mentioned support information may be information presented to the target individual, or it may be intermediate information used to generate the information presented to the target individual. For example, the target individual's current income, expenses, assets, and their breakdown can be used for the target individual's life planning and can be generated using data from predetermined data items (e.g., income and expense data). For this reason, the support information generation unit 102 may generate support information that shows such amounts.

[0023] For example, when generating support information that shows the breakdown of expenditures, the classification unit 101 uses numerical data indicating the expenditure amount for each expenditure of the subject and content information indicating the content of each expenditure to classify the numerical data into data items corresponding to each breakdown of expenditures (for example, housing expenses, food expenses, etc.). Then, the support information generation unit 102 uses the numerical data indicating the expenditure amount for each expenditure of the subject as data corresponding to the breakdown classified by the classification unit 101 to generate support information that shows the breakdown of the subject's expenditures.

[0024] As described above, the information processing device 1 according to this exemplary embodiment employs a configuration that includes a classification unit 101 that classifies multiple numerical data related to the subject's life plan into data items for generating support information to support the subject's life planning, using content information indicating the content of each numerical data and a language model that has been machine-learned from natural language, and a support information generation unit 102 that generates support information using the numerical data as data for the data items classified by the classification unit 101.

[0025] According to the above configuration, multiple numerical data related to life plans are classified into predetermined data items without human intervention, and support information is automatically generated based on the classification results. Therefore, the above configuration has the effect of reducing the burden of manual work in life planning.

[0026] Furthermore, the information processing device 1 can also support the decision-making of the subject by generating support information. The information processing device 1 can also be used in healthcare. For example, the information processing device 1 can generate support information showing the trends in medical expenses. By presenting such support information to the subject, it can encourage them to be more mindful of their health management.

[0027] (Support Program) The functions of the information processing device 1 described above can also be implemented by a program. The support program according to this exemplary embodiment is a life planning support program in which a computer functions as a classification means for classifying multiple numerical data related to a person's life plan into data items for generating support information to support the person's life planning, using content information indicating the content of each numerical data and a language model that has been trained on natural language; and a support information generation means for generating support information using the numerical data as data for the data items classified by the classification means. This support program has the effect of reducing the burden of manual work in life planning.

[0028] (Flow of support methods) The flow of the support method according to this exemplary embodiment will be explained with reference to Figure 2. Figure 2 is a flowchart showing the flow of the support method. Note that the entity executing each step in this support method may be a processor provided in the information processing device 1, a processor provided in another device, or the entity executing each step may be a processor provided in a different device.

[0029] In S1 (classification process), at least one processor classifies multiple numerical data points related to the subject's life plan into data items that generate support information to assist the subject's life planning, using content information that describes the content of each numerical data point and a language model that has been trained on natural language.

[0030] In S2 (support information generation process), at least one processor uses the numerical data described above as data for the data items classified in the classification process of S1 to generate support information.

[0031] As described above, the support method according to this exemplary embodiment is a life planning support method, and employs a configuration in which at least one processor classifies multiple numerical data related to the subject's life plan into data items for generating support information to support the subject's life planning, using content information indicating the content of each numerical data and a language model that has been machine-learned to learn natural language; and a support information generation process that generates support information using the numerical data as data for the data items classified in the classification process. Therefore, the support method according to this embodiment has the effect of reducing the burden of manual work in life planning.

[0032] [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.

[0033] (Overview of Information Processing Device 1A) An overview of the information processing device 1A according to this exemplary embodiment will be described with reference to Figure 3. Figure 3 is a diagram showing an example of processing performed by the information processing device 1A.

[0034] In the example shown in Figure 3, the subject is inputting a deposit / withdrawal statement 301 into the information processing device 1A. The deposit / withdrawal statement 301 shows the amount of deposits and withdrawals in the subject's account, as well as related information such as the date of the deposit or withdrawal and a string indicating the content of the deposit or withdrawal. The information processing device 1A acquires the above amount shown in the deposit / withdrawal statement 301 as numerical data, and acquires the above string as content information indicating the content of the numerical data.

[0035] Next, the information processing device 1A uses the acquired content information and a language model M trained on natural language to classify the acquired numerical data into data items for generating support information to assist the subject's life planning. Specifically, the information processing device 1A generates a prompt 302 that includes the acquired content information and the data items, instructing the language model M to infer the relationship between the numerical data and the data items. The information processing device 1A then inputs the generated prompt 302 into the language model M to output an inference result 303. In the example in Figure 3, the inference result 303 shows which of the predetermined data items (which can also be called classification categories) in income and expenditure each of the deposit and withdrawal items shown in the deposit and withdrawal statement 301 corresponds to.

[0036] Next, the information processing device 1A generates support information 304 using the inference result 303 and presents the generated support information 304 to the subject. In the example in Figure 3, the support information 304 shows the predicted income and expenditure amounts of the subject at age 65, and their breakdown. The support information 304 is generated using the income and expenditure amounts of each data item shown in the inference result 303.

[0037] As described above, with the information processing device 1A, the user can simply input deposit and withdrawal details 301 into the information processing device 1A, receive support information 304, and use this information to create their own life plan. In this way, the information processing device 1A can reduce the burden of manual work in life planning.

[0038] (Configuration of Information Processing Device 1A) The configuration of the information processing device 1A according to this exemplary embodiment will be described with reference to Figure 4. Figure 4 is a block diagram showing the configuration of the information processing device 1A. The information processing device 1A is a device equipped with functions to support life planning. The information processing device 1A may be a local device used by individual users, or it may be a server that provides life planning support services to multiple users.

[0039] As shown in the figure, the information processing device 1A includes a control unit 10A that controls all parts of the information processing device 1A, and a storage unit 11A that stores various data used by the information processing device 1A. The information processing device 1A also includes a communication unit 12A for the information processing device 1A to communicate with other devices, an input unit 13A that receives input to the information processing device 1A, and an output unit 14A for the information processing device 1A to output data. The control unit 10A includes a classification unit 101A, a support information generation unit 102A, a data acquisition unit 103A, a reception unit 104A, a presentation control unit 105A, and an aggregation unit 106A.

[0040] The classification unit 101A, similar to the classification unit 101 in Exemplary Embodiment 1, classifies multiple numerical data related to the subject's life plan into data items for generating support information to assist the subject's life planning, using content information indicating the content of each numerical data and a language model M trained on natural language. In this exemplary embodiment, an example is described in which the language model M accepts input of a text-formatted prompt written in natural language and outputs a response in natural language. However, the language model M may also be a model capable of accepting input of data in formats other than text data, such as images.

[0041] The support information generation unit 102A generates support information using numerical data related to the subject's life plan as data for data items classified by the classification unit 101, similar to the support information generation unit 102 in exemplary embodiment 1. The support information generated by the support information generation unit 102A is not limited to information like the support information 304 shown in Figure 3, and only needs to directly or indirectly support the subject's life planning.

[0042] The data acquisition unit 103A acquires various data necessary to support the life planning of the target person. For example, the data acquisition unit 103A acquires numerical data that is subject to classification by the classification unit 101A, as well as content information indicating the content of that data. The method of data acquisition by the data acquisition unit 103A is arbitrary. For example, the data acquisition unit 103A may acquire data input via the input unit 13A, or it may acquire data from other devices via the communication unit 12A. Alternatively, for example, the data acquisition unit 103A may acquire numerical data and content information by performing character recognition on image data scanned from receipts, etc. Character recognition may be performed by a device other than the information processing device 1A.

[0043] The reception unit 104A receives various instructions related to supporting the life planning of the subject. For example, the reception unit 104A receives instructions to modify the results of inference by the language model M. The method of receiving instructions is arbitrary. For example, the reception unit 104A may receive instructions input via the input unit 13A, or it may receive instructions from other devices via the communication unit 12A.

[0044] The presentation control unit 105A presents various information related to supporting the life planning of the target person. For example, the presentation control unit 105A presents support information generated by the support information generation unit 102A. Also, for example, the presentation control unit 105A presents the inference results output by the language model M. As will be described in detail later, the presentation control unit 105A may present the above inference results along with the basis for the inference. The method of presenting the information is arbitrary. For example, the presentation control unit 105A may present the information by having the output unit 14A output the information, or by having the communication unit 12A output the information to another device. Furthermore, the information can be presented in any manner, such as display, printing, voice, or a combination thereof.

[0045] The aggregation unit 106A aggregates multiple numerical data acquired by the data acquisition unit 103A according to the data items classified by the classification unit 101A. As will be described in detail later, the results of the aggregation by the aggregation unit 106A are used to generate support information by the support information generation unit 102.

[0046] The aggregation method used by the aggregation unit 106A may be one that suits the intended use of the aggregation results. Furthermore, the aggregation unit 106A may aggregate numerical data using different methods for each data item. For example, when generating support information showing expenditure amounts for each predetermined data item, the data acquisition unit 103A may aggregate these amounts by summing up the amounts classified under the same data item. Also, for example, when generating support information showing average expenditure amounts over a predetermined period, the aggregation unit 106A may calculate the average expenditure amount over the above period.

[0047] As described above, the information processing device 1A includes a classification unit 101A that classifies multiple numerical data related to the subject's life plan into data items for generating support information to assist the subject's life planning, using content information indicating the content of each numerical data and a language model M that has been machine-learned from natural language; and a support information generation unit 102A that generates support information using the numerical data related to the subject's life plan as data for the data items classified by the classification unit 101A. Therefore, similar to the information processing device 1, the effect of reducing the burden of manual work in life planning can be obtained.

[0048] Furthermore, as described above, the information processing device 1A includes an aggregation unit 106A that aggregates multiple numerical data for each data item classified by the classification unit 101A, and the support information generation unit 102A generates support information using the aggregation results from the aggregation unit 106A. As a result, in addition to the effects of the information processing device 1, it is possible to generate support information that reflects the aggregation results of numerical data without having to perform the cumbersome task of aggregating numerical data for each data item manually.

[0049] (Example of prompt and inference result) Figure 5 shows a prompt generated by the classification unit 101A and an example of the inference result obtained by inputting that prompt into the language model M.

[0050] The prompt 501 shown in Figure 5 instructs the system to infer the relationship between each item in the deposit / withdrawal statement (all of which include numerical data) and the "income / expense category," a data item used to generate support information. The prompt 501 also includes the statement, "You are a financial planner creating a life plan." While including such a statement is not mandatory, it is expected to improve the accuracy of the inference.

[0051] Furthermore, prompt 501 instructs the system to check whether it can infer the income / expense category for each item in the deposit / withdrawal statement. The wording of the prompt can be changed as appropriate to obtain the desired inference result. For example, the classification unit 101A may generate prompts with different inference instructions depending on the target numerical data, content information, data items to be classified, and the language model used.

[0052] Furthermore, prompt 501 includes an "answer format" and instructs the user to answer using this format. By specifying the answer format in this way, it becomes possible to obtain inference results in a format that is easy to use for generating support information. This answer format also includes an item called "basis for inference." By including such an item, the language model M can output the basis for the inference along with the inference result. Note that the instruction to output the basis for the inference may also be given by including a sentence such as "Please answer with the basis along with the inference result" in the prompt.

[0053] Furthermore, prompt 501 may include sentences specifying output conditions. For example, prompt 501 may include sentences such as, "The income / expense category must be inferred for all items in the deposit / expense statement," or "The number of elements included in the response format must match the number of items in the deposit / expense statement." This makes it possible to improve the accuracy of the output of the language model M.

[0054] In prompt 501, everything except the "deposit and withdrawal details" is standard. For this reason, everything except the "deposit and withdrawal details" can be stored as a standard template in the storage unit 11A or the like. This allows the classification unit 101A to input the numerical data (indicating the deposit and withdrawal amounts) and content information (indicating the details of the deposit and withdrawal) shown in the deposit and withdrawal details acquired by the data acquisition unit 103A into the above template and generate prompt 501.

[0055] The inference result 502 shown in Figure 5 is an example of an inference result obtained by inputting the prompt 501 into the language model M. The inference result 502 shows the result of inferring the relationship between each item in the deposit and withdrawal statement (all of which include numerical data) and the "income / expense category," which is a data item used to generate support information, in the format of the response format shown in the prompt 501. Specifically, in the inference result 502, each item in the deposit and withdrawal statement is associated with the "income / expense category" to which that item belongs. Therefore, the classification unit 101A can classify each numerical data based on the inference result 502.

[0056] Note that while prompt 501 in Figure 5 is a prompt for classifying multiple numerical data at once, classification may also be performed on individual numerical data. In this case, the classification unit 101A can input the numerical data and one of the target data items into the language model M for each numerical data item, and then determine whether the numerical data corresponds to that data item. This process can be performed for each numerical data item. Alternatively, the classification unit 101A can input the numerical data and each of the target data items into the language model M, and then determine which of the data items the numerical data corresponds to.

[0057] Furthermore, the classification unit 101A may input one of the data items to be classified and one of the numerical data into the language model M, and perform the process for each of the multiple numerical data to determine whether the numerical data corresponds to that data item. By performing this process for each data item, each numerical data can be classified into the corresponding data item. Alternatively, the classification unit 101A may input one data item and multiple numerical data into the language model M, and perform the process for each of the multiple data items to determine which of the data items the numerical data corresponds to. In either case, content information indicating the content of the numerical data is input into the language model M.

[0058] The format in which the inference results are output can be specified by a prompt. For example, the classification unit 101A may generate a prompt instructing the user to answer with one of three options: belongs to a data item, neutral, or does not belong to a data item. In this case, when the classification unit 101A outputs a response indicating that a numerical data item belongs to a data item, it should classify that numerical data item. The processing to be performed when a neutral response is output can be predetermined. For example, when the presentation control unit 105A outputs a neutral response to a prompt asking whether a numerical data item belongs to a data item, it may present the user with combinations of numerical data and data items and ask them to input whether the numerical data belongs to that data item. Alternatively, for example, the classification unit 101A may generate a prompt instructing the user to output a numerical value (for example, a number between 0 and 1) indicating the likelihood of belonging to a data item. In this case, the classification unit 101A should classify the numerical data into a data item for which the output numerical value is above a predetermined threshold.

[0059] Furthermore, the reasoning result 502 also shows the basis for the reasoning. As will be explained in detail later based on Figure 6, the basis for the reasoning is presented to the subject by the presentation control unit 105A as a basis for judging whether the reasoning result is appropriate or not.

[0060] As described above, the classification unit 101A includes content information indicating the content of numerical data and data items for generating support information, generates a prompt instructing the system to infer the relationship between the numerical data and the data items, and may classify the numerical data based on the output obtained by inputting the generated prompt to the language model M. This makes it possible to appropriately classify each numerical data.

[0061] (Regarding the presentation and reclassification of inference results) The inference results of language model M are not always correct. Therefore, it may be helpful to present the inference results of language model M to the subjects and ask them to check for errors. If errors are found, the subjects may be asked to correct them, and the numerical data may be reclassified to reflect these corrections. This will be explained using Figure 6. Figure 6 is a diagram illustrating the presentation of inference results and reclassification.

[0062] Screen example 601 in Figure 6 is an example of a UI (User Interface) screen that presents inference results and accepts correction instructions. Screen example 601 shows the inference result 502 shown in Figure 5, and prompts the user to check if there are any errors in the inference result. If there are errors, it asks the user to input the details and instruct the system to re-output the result. Screen example 601 also displays a text box for entering correction details and a button (software key) for instructing the system to re-output the result.

[0063] The presentation control unit 105A can allow the subject to confirm the inference results by presenting a UI screen, such as the example screen 601. The reception unit 104A can also accept corrections to the inference results via the UI screen, such as the example screen 601. If the output unit 14A has the function of displaying and outputting an image, the presentation control unit 105A may display the UI screen on the output unit 14A. Alternatively, the presentation control unit 105A may display the UI screen on an external display device of the information processing device 1A (for example, a display device on the terminal device used by the subject) via the communication unit 12A.

[0064] In the example screen 601, when the correction content is entered into the text box and the button to instruct re-output is operated, the receiving unit 104A accepts the entered correction content as a correction instruction. For example, the receiving unit 104A may accept a correction instruction that includes the correct income / expense category and the reason why the inference result is incorrect. In that case, the presentation control unit 105A may display a pull-down list listing each of the income / expense categories and allow the user to select the correct income / expense category.

[0065] Then, the classification unit 101A reclassifies the numerical data based on the correction instructions received by the reception unit 104A. For example, the classification unit 101A may generate the prompt 602 shown in Figure 6 and input it to the language model M, and then reclassify the numerical data based on the reinference result 603 output from the language model M.

[0066] Prompt 602 is a prompt that includes the correction instructions received by the reception unit 104A and instructs the system to infer the relationship between numerical data and data items based on those correction instructions. Prompt 602 also includes the sentence, "However, do not output any text that is not included in the answer format." Thus, even in prompts that instruct re-inference, sentences specifying output conditions may be included. This prevents unnecessary content (for example, sentences such as "Apologies. I will re-output." or "Content has been updated") from being output from the language model M, thus preventing any disruption to the generation of support information. Prompt 602, like prompt 501 shown in Figure 5, can be generated using a predetermined template. In addition, prompt 602 may also include deposit and withdrawal details, income and expenditure categories, and an answer format, similar to prompt 501.

[0067] The reinference result 603 shown in Figure 6 differs from the inference result 502 shown in Figure 5 in that the "category" of "Store A" has been changed to "clothing expenses," and the basis for the inference has been changed to the sentence "Store A is a clothing store." In this way, by performing reinference using the prompt 602 which includes the natural language sentence received as a correction instruction, it is possible to output a reinference result 603 that reflects the content of the correction instruction.

[0068] As described above, the classification unit 101A may generate a prompt instructing the system to output the reasoning results for the relationship between numerical data relating to the life planning target and data items for generating support information to assist the target's life planning, along with the basis for the reasoning. The presentation control unit 105A may then present the reasoning results output by the language model M along with the reasoning. This provides the effect that, in addition to the effects of the information processing device 1, the system can consider the appropriateness of the reasoning results of the language model M using the reasoning as a basis for judgment. The presentation control unit 105A may also present the classification results of the classification unit 101A instead of the reasoning results output by the language model M.

[0069] Furthermore, as described above, the information processing device 1A is equipped with a receiving unit 104A that receives correction instructions for the inference results. The classification unit 101A then reclassifies the numerical data based on the correction instructions received by the receiving unit 104A. As a result, in addition to the effects performed by the information processing device 1, it becomes possible to correct errors in the inference results and generate appropriate support information.

[0070] Furthermore, as described above, the reception unit 104A may accept correction instructions expressed in natural language. In this case, the classification unit 101A generates a prompt that includes the correction instructions received by the reception unit 104A and instructs it to infer the relationship between numerical data and data items based on those instructions. The classification unit then inputs the generated prompt into the language model M and reclassifies the numerical data based on the output obtained. This provides the added benefit of enabling appropriate reclassification in accordance with the intent of the correction instructions, in addition to the effects of the information processing device 1.

[0071] For example, even if the correction instruction comment in the example in Figure 6 does not include the correct category "clothing expenses," such as "Store A is a clothing store" or "What I bought at Store A was a shirt," it is still possible to infer the correct category "clothing expenses" by considering the content of the comment.

[0072] Furthermore, by accepting correction instructions expressed in natural language, it becomes possible to absorb inconsistencies in spelling and correct similar items in bulk. For example, suppose that for the aforementioned "Store A," the bank account statement shows it as "A-Shoten," while the credit card statement shows it as "A-Shoten." In such a case, if a correction instruction is entered stating that "Store A" is a clothing store, it will be possible to classify both the expenditure to "A-Shoten" in the bank account statement and the expenditure to "A-Shoten" in the credit card statement as "Clothing Expenses." Also, for example, based on a correction instruction for "Store A," it is possible to correct the classification of multiple stores with the same name, such as "Store A X Branch," "Store A Y Branch," etc., in bulk.

[0073] (Reuse of correction instructions) The content of the correction instructions described above may be recorded in the memory unit 11A or an external database, etc., and used for subsequent inference. For example, the comment of the correction instruction entered when classifying certain numerical data, "Store A is a clothing store. Therefore, the category is clothing expenses, not food expenses," may be recorded, and this comment may be used for classifying other numerical data as well. This will enable the correct classification of "Store A" as "clothing expenses" in subsequent classifications (which may be limited to classifications of the same subject, or may be reused for classifications of other subjects).

[0074] Furthermore, the presentation control unit 105A may present the recorded correction instructions to the subject, and the reception unit 104A may accept corrections, deletions, and additions to the correction instructions. This allows the content of correction instructions that align with the subject's intentions to be reflected in subsequent classifications without requiring retraining of the language model M. The correction instructions can be reused in the same way as when they are used for the first time, by including them in the prompt. In other words, the classification unit 101A can include the recorded correction instructions, generate a prompt that instructs the language model M to infer the relationship between numerical data and data items based on the correction instructions, and classify the numerical data based on the output obtained by inputting the generated prompt into the language model M.

[0075] (Process flow) The processing flow performed by the information processing device 1A will be explained with reference to Figure 7. Figure 7 is a flowchart showing the processing flow performed by the information processing device 1A. The flowchart in Figure 7 includes each process of the support method according to this exemplary embodiment.

[0076] In S11, the data acquisition unit 103A acquires numerical data to be classified by the classification unit 101A, along with content information indicating its content. For example, the data acquisition unit 103A may acquire the deposit and withdrawal details of the subject (including numerical data and content information) via the input unit 13A or the communication unit 12A.

[0077] In S12, the classification unit 101A generates a prompt to be input to the language model M. Specifically, the classification unit 101A generates a prompt that includes numerical data and content information acquired in S11, and predetermined data items for generating support information to assist the subject's life planning, and instructs the model to infer the relationship between the numerical data and the data items.

[0078] In S13, the classification unit 101A inputs the prompt generated in S12 into the language model M to infer the relationship between numerical data and data items. In S14, the presentation control unit 105A presents the inference result from S13 to the subject. In S15, the reception unit 104A determines whether or not there are any correction instructions for the inference result presented in S14.

[0079] In S14, the presentation control unit 105A may present the inference result by displaying a UI screen, such as the example screen 601 in Figure 6. In that case, the reception unit 104A may receive a correction instruction via the UI screen. If the result in S15 is determined to be YES, the process proceeds to S16; if the result in S15 is determined to be NO, the process proceeds to S18.

[0080] In S16, the reception unit 104A records the contents of the received correction instructions in the storage unit 11A or an external database, etc. The timing of recording the contents of the correction instructions is arbitrary. For example, the reception unit 104A may record the contents of the correction instructions after the classification described later has been completed, or after the support information has been presented.

[0081] In S17, the classification unit 101A generates a prompt that reflects the content of the correction instruction received by the reception unit 104A, specifically a prompt that includes the correction instruction and instructs the system to infer the relationship between numerical data and data items based on the correction instruction. After this, the process returns to S13, where the classification unit 101A inputs the newly generated prompt into the language model M to infer the relationship between numerical data and data items.

[0082] In S18, the classification unit 101A classifies the numerical data acquired in S11 based on the inference result of the language model M (or the latest inference result if multiple inferences were performed). In the flowchart of Figure 7, the processes in S12, S13, S17, and S18 correspond to the classification process that classifies multiple numerical data related to the subject's life plan into data items for generating support information to assist the subject's life planning, using content information indicating the content of each numerical data and a language model trained on natural language.

[0083] In S19, the aggregation unit 106A aggregates the multiple numerical data acquired in S11 according to the data items classified in S18. In other words, the aggregation unit 106A aggregates multiple numerical data classified under the same data item.

[0084] In S20 (support information generation process), the support information generation unit 102A generates support information using the numerical data acquired in S11 as data for the data items classified in S18. At this time, for data items that have been aggregated in S19, the support information generation unit 102A uses the aggregation results as data for those data items.

[0085] In S21, the presentation control unit 105A presents the support information generated in S20 to the target person. This completes the process shown in Figure 7.

[0086] (Regarding the information used to generate support information) In addition to the numerical data mentioned above, it is possible to generate more comprehensive support information by using various types of information other than numerical data. For example, by using information such as family structure and future plans, the support information generation unit 102A can generate support information that shows the results of a simulation based on that information.

[0087] Furthermore, each piece of information used to generate support information, including numerical data, may be obtained through dialogue with the subject using a language model M. In this case, the information processing device 1A simply needs to input text data representing the content of the subject's utterances into the language model M, generate a response to those utterances, and then present the generated response to the subject, repeating this process. In addition, when generating the response, the accuracy of the response can be improved by having the language model M refer to reliable materials related to life planning (for example, statistical information and reports provided and accumulated by insurance companies or the government). Furthermore, when the information processing device 1A has the language model M generate a response to a question, it may also output the basis for that response and present the outputted response along with its basis to the subject.

[0088] If, for example, the subject's hobbies can be elicited through the dialogue described above, the support information generation unit 102A can generate support information that reflects those hobbies. For example, if the subject's hobby is riding motorcycles, the unit can generate support information that includes simulation results of what would happen if a motorcycle accident occurred, as well as support information that includes recommendations for insurance policies that provide generous compensation for injuries and property damage resulting from motorcycle accidents.

[0089] Furthermore, dialogue with the target individual can also take place after the support information has been presented. For example, the information processing device 1A may ask the target individual to input their impressions and questions regarding the presented support information, and then have the language model M generate and present answers to these inputs. This can deepen the target individual's understanding of the support information and reveal their values ​​and way of thinking. The answers and questions presented to the target individual can also be determined using a rule-based system.

[0090] Furthermore, the support information generation unit 102A may update the support information using information obtained after presenting the support information, or it may make new suggestions (for example, suggestions for insurance plans or financial products) using information obtained after presenting the support information. In this case, the information processing device 1A may use the language model M to generate a summary of the subject's values ​​and way of thinking from the subject's utterances and present it to the subject as reference information. In addition, in this case, each sentence included in the summary may be presented in association with the subject's utterances, making it easy to confirm whether the summary is in line with the subject's intentions.

[0091] Furthermore, users can have their support information recreated at any time. They can also generate multiple sets of support information by changing the input information (for example, income and its rate of change, retirement age, number of children and educational destinations, etc.). This allows users to simulate life plans that are tailored to their income and changes in it.

[0092] Furthermore, the information processing device 1A can also recommend changes to insurance and other policies in accordance with changes in the subject's life plan. It should be noted that the support information generated by the information processing device 1A is for reference only and is not necessarily accurate. Therefore, in addition to presenting the generated support information, the information processing device 1A may also provide information introducing financial planners and insurance sales representatives, or accept reservations for meetings with these individuals.

[0093] Furthermore, the information processing device 1A may start monitoring the subject's income and expenses after generating support information that includes the simulation results of the subject's future income and expenses. As a result, when a discrepancy occurs between the simulation results shown in the previously generated support information and the actual income and expenses, the support information generation unit 102A generates support information indicating this discrepancy, and the presentation control unit 105A can then present the generated support information to the subject.

[0094] This allows individuals to recognize that they are in a situation where a review of their life plan is necessary, and to consider taking action early. Information regarding the individual's income and expenses (e.g., changes in salary income, withdrawal amounts, investment returns, loan balances, interest rate changes, etc.) can be obtained by linking with, for example, the individual's bank account, securities account, or credit card system. In particular, since the information processing device 1A can automatically classify numerical data, it can automatically and accurately detect situations where a review of the life plan is necessary.

[0095] [Variation] The entities executing each process described in the exemplary embodiments above are arbitrary and not limited to the examples given. For example, a system having the same functions as the information processing devices 1 and 1A can be constructed using multiple devices that can communicate with each other. Furthermore, the entities executing each process shown in the flowchart in Figure 7 may be a single device (which can also be called a processor) or multiple devices (which can also be called processors).

[0096] [Examples of implementation using software] Some or all of the functions of the information processing devices 1,1A (hereinafter also referred to as "each of the above devices") may be implemented by hardware such as integrated circuits (IC chips) or by software.

[0097] 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.

[0098] Computer C comprises at least one processor C1 and at least one memory C2. Memory C2 stores a program (support 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 devices.

[0099] 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.

[0100] 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.

[0101] 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.

[0102] 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.

[0103] [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.

[0104] (Note A1) An information processing device comprising: a classification means for classifying multiple numerical data related to a subject's life plan into data items for generating support information to support the subject's life planning, using content information indicating the content of each numerical data and a language model that has been machine-learned to natural language; and a support information generation means for generating the support information using the numerical data as data for the data items classified by the classification means.

[0105] (Appendix A2) The information processing apparatus described in Appendix A1, comprising an aggregation means for aggregating the plurality of numerical data for each data item classified by the classification means, and the support information generation means for generating the support information using the results of the aggregation by the aggregation means.

[0106] (Note A3) The information processing apparatus according to Appendix A1 or A2, wherein the classification means includes the content information and the data items, generates a prompt instructing the system to infer the relationship between the numerical data and the data items, and classifies the numerical data based on the output obtained by inputting the generated prompt to the language model.

[0107] (Note A4) The information processing apparatus according to Appendix A3, wherein the classification means generates a prompt instructing the output of the basis for the inference along with the inference result of the relationship between the numerical data and the data item, and the presentation control means presents the inference result output by the language model along with the basis.

[0108] (Note A5) The information processing apparatus described in Appendix A4, comprising a receiving means for receiving instructions for correction of the inference results, wherein the classification means reclassifies the numerical data based on the correction instructions received by the receiving means.

[0109] (Note A6) The information processing apparatus described in Appendix A5, wherein the receiving means receives a correction instruction expressed in natural language, the classification means generates a prompt that includes the correction instruction received by the receiving means and instructs the classification means to infer the relationship between the numerical data and the data item based on the correction instruction, and the numerical data is reclassified based on the output obtained by inputting the generated prompt into the language model.

[0110] (Note B1) A support method comprising: a classification process in which at least one processor classifies multiple numerical data relating to a subject's life plan into data items for generating support information to support the subject's life planning, using content information indicating the content of each numerical data and a language model trained on natural language; and a support information generation process that generates the support information using the numerical data as data for the data items classified in the classification process.

[0111] (Note B2) The support method according to Appendix B1, wherein the at least one processor includes an aggregation process that aggregates the plurality of numerical data for each data item classified in the classification process, and in the support information generation process, the at least one processor generates the support information using the results of the aggregation by the aggregation process.

[0112] (Note B3) The support method according to Appendix B1 or B2, wherein in the classification process, the at least one processor generates a prompt that includes the content information and the data items and instructs the processor to infer the relationship between the numerical data and the data items, and classifies the numerical data based on the output obtained by inputting the generated prompt to the language model.

[0113] (Note B4) The support method according to Appendix B3, wherein in the classification process, the at least one processor generates a prompt instructing the at least one processor to output the basis for the inference along with the inference result of the relationship between the numerical data and the data item, and the at least one processor presents the inference result output by the language model along with the basis.

[0114] (Note B5) The support method according to Appendix B4, wherein the at least one processor performs an acceptance process to receive correction instructions for the inference result, and the at least one processor reclassifies the numerical data based on the correction instructions received in the acceptance process.

[0115] (Note B6) The support method described in Appendix B5, wherein in the reception process, at least one processor receives a correction instruction expressing the content of the correction in natural language, the at least one processor generates a prompt that includes the correction instruction received in the reception process and instructs the system to infer the relationship between the numerical data and the data item based on the correction instruction, and the system reclassifies the numerical data based on the output obtained by inputting the generated prompt into the language model.

[0116] (Note C1) A support program that causes a computer to function as a classification means for classifying multiple numerical data related to a subject's life plan into data items for generating support information to assist the subject's life planning, using content information indicating the content of each numerical data and a language model trained on natural language. A support program that generates the support information by using the numerical data as data for the data items classified by the classification means.

[0117] (Note C2) The support program described in Appendix C1, wherein the computer functions as an aggregation means for aggregating the plurality of numerical data for each data item classified by the classification means, and the support information generation means generates the support information using the results of the aggregation by the aggregation means.

[0118] (Note C3) The classification means includes the content information and the data items, generates prompts instructing the system to infer the relationship between the numerical data and the data items, and classifies the numerical data based on the output obtained by inputting the generated prompts into the language model, as described in Appendix C1 or C2.

[0119] (Note C4) The support program described in Appendix C3, wherein the classification means generates a prompt instructing the computer to output the basis for the inference along with the inference result of the relationship between the numerical data and the data item, and causes the computer to function as a presentation control means that presents the inference result output by the language model along with the basis.

[0120] (Note C5) The support program described in Appendix C4, wherein the computer functions as a receiving means for receiving instructions to correct the inference results, and the classification means reclassifies the numerical data based on the correction instructions received by the receiving means.

[0121] (Appendix C6) The support program described in Appendix C5, wherein the receiving means receives a correction instruction expressed in natural language, the classification means generates a prompt that includes the correction instruction received by the receiving means and instructs the classification means to infer the relationship between the numerical data and the data item based on the correction instruction, and the numerical data is reclassified based on the output obtained by inputting the generated prompt into the language model.

[0122] (Note D1) An information processing device comprising at least one processor, wherein the at least one processor performs a classification process that classifies multiple numerical data relating to a subject's life plan into data items for generating support information to support the subject's life planning, using content information indicating the content of each numerical data and a language model trained on natural language. The information processing device also performs a support information generation process that generates the support information using the numerical data as data for the data items classified in the classification process.

[0123] The information processing 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.

[0124] (Note D2) The information processing apparatus according to Appendix D1, wherein the at least one processor performs an aggregation process to aggregate the plurality of numerical data for each data item classified by the classification process, and in the support information generation process, the at least one processor generates the support information using the aggregation results of the aggregation process.

[0125] (Note D3) The information processing apparatus according to Appendix D1 or D2, wherein in the classification process, the at least one processor generates a prompt that includes the content information and the data items and instructs the system to infer the relationship between the numerical data and the data items, and classifies the numerical data based on the output obtained by inputting the generated prompt to the language model.

[0126] (Note D4) The information processing apparatus according to Appendix D3, wherein in the classification process, the at least one processor generates a prompt instructing the output of the basis for the inference along with the inference result of the relationship between the numerical data and the data item, and the at least one processor performs a presentation control process to present the inference result output by the language model along with the basis.

[0127] (Note D5) The information processing apparatus according to Appendix D4, wherein the at least one processor performs an acceptance process to receive a correction instruction for the inference result, and the at least one processor reclassifies the numerical data based on the correction instruction received in the acceptance process.

[0128] (Note D6) The information processing apparatus according to Appendix D5, wherein in the reception process, at least one processor receives a correction instruction expressing the content of the correction in natural language, the at least one processor generates a prompt that includes the correction instruction received in the reception process and instructs the system to infer the relationship between the numerical data and the data item based on the correction instruction, and reclassifies the numerical data based on the output obtained by inputting the generated prompt into the language model.

[0129] (Note E) A non-temporary recording medium that records a support program for causing a computer to function as an information processing device, the support program which causes the computer to execute a classification process that classifies multiple numerical data related to a subject's life plan into data items for generating support information to support the subject's life planning, using content information indicating the content of each numerical data and a language model that has been machine-learned from natural language, and a support information generation process that generates the support information using the numerical data as data for the data items classified by the classification process. [Explanation of Symbols]

[0130] 1. Information Processing Device 101 Classification section (classification means) 102 Support information generation unit (support information generation means) 1A Information Processing Device 101A Classification section (classification means) 102A Support information generation unit (support information generation means) 104A Reception area (reception method) 105A Presentation Control Unit (Presentation Control Means) 106A Aggregation Unit (Aggregation Means) M language model

Claims

1. A classification means that classifies multiple numerical data related to the subject's life plan into data items for generating support information to assist the subject's life planning, using content information that indicates the content of each numerical data and a language model that has been trained on natural language. An information processing apparatus comprising: a support information generation means that generates the support information by using the numerical data as data for data items classified by the classification means.

2. The system includes an aggregation means for aggregating the plurality of numerical data for each data item classified by the classification means, The information processing apparatus according to claim 1, wherein the support information generation means generates the support information using the results of aggregation by the aggregation means.

3. The information processing apparatus according to claim 1 or 2, wherein the classification means includes the content information and the data items, generates a prompt instructing the language model to infer the relationship between the numerical data and the data items, and classifies the numerical data based on the output obtained by inputting the generated prompt to the language model.

4. The classification means generates a prompt that instructs the system to output the inference result of the relationship between the numerical data and the data item, along with the basis for the inference. The information processing apparatus according to claim 3, further comprising presentation control means for presenting the inference result output by the language model or the classification result of the classification means together with the basis.

5. The system includes a receiving means for receiving instructions to modify the aforementioned inference results, The information processing apparatus according to claim 4, wherein the classification means reclassifies the numerical data based on the correction instructions received by the receiving means.

6. The aforementioned receiving means receives correction instructions that express the content of the correction in natural language, The information processing apparatus according to claim 5, wherein the classification means includes a correction instruction received by the receiving means, generates a prompt instructing the system to infer the relationship between the numerical data and the data item based on the correction instruction, and reclassifies the numerical data based on the output obtained by inputting the generated prompt into the language model.

7. At least one processor, A classification process that classifies multiple numerical data related to the subject's life plan into data items for generating support information to assist the subject's life planning, using content information that indicates the content of each numerical data and a language model that has been trained on natural language. A support method that performs a support information generation process, which generates the support information by using the numerical data as data for data items classified in the classification process.

8. Computers, A classification means that classifies multiple numerical data related to the subject's life plan into data items for generating support information to assist the subject's life planning, using content information that indicates the content of each numerical data and a language model that has been trained on natural language. A support program that functions as a support information generation means, which generates the support information by using the numerical data as data for data items classified by the classification means.

Images