Information processing device, information processing method, and program

Abstract

A technology is provided for reducing the effort required by users to monitor data to be analyzed. [Solution] An information processing device (1) according to one aspect of the present disclosure is configured to monitor whether a result (30) of a first analysis of continuously collected target data (2) satisfies a notification condition (4), and, if the result (30) of the first analysis satisfies the notification condition (4) as a result of the monitoring, notify a user terminal (U0) of notification data (5) indicating a result (35) of a second analysis of the target data (2).

Description

[Technical Field]

[0001] The present disclosure relates to an information processing device, an information processing method, and a program. [Background technology]

[0002] Data analysis results are utilized in various fields, including retail, advertising, sensing, finance, and payment. For example, in the retail industry, data analysis results are utilized to achieve user-desired goals in all fields, such as identifying best-selling products. In recent years, various systems have been proposed for providing such data analysis results. For example, Patent Document 1 proposes an analysis support system for executing analysis processing triggered by a user. Specifically, the proposed analysis support system is configured to receive analysis instructions from a user terminal, instruct an external information processing system to execute analysis based on the received analysis instructions, and output the obtained analysis results. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 7540808 Summary of the Invention [Problem to be solved by the invention]

[0004] According to conventional methods such as those disclosed in Patent Document 1, desired analysis results can be obtained in response to instructions from a user. However, the present inventors have discovered the following problem with these conventional methods. That is, in conventional methods, data analysis is performed when a user issues instructions to a system. Therefore, in order to issue analysis instructions to the system at an appropriate timing according to the purpose, the user must monitor the data to be analyzed and determine the timing to issue the analysis instructions based on the monitoring results. Monitoring this data is time-consuming.

[0005] In one aspect, the present disclosure has been made in consideration of the above circumstances, and one of its objectives is to provide a technology for reducing the effort required by users to monitor data to be analyzed. [Means for solving the problem]

[0006] In order to solve the above-mentioned problems, the present disclosure employs the following configurations. Note that the following configurations can be combined as appropriate.

[0007] An information processing device according to one aspect of the present disclosure includes a control unit configured to monitor whether a result of a first analysis of continuously collected target data satisfies a notification condition, and, if the result of the first analysis satisfies the notification condition as a result of the monitoring, notify a user terminal of notification data indicating a result of a second analysis of the target data.

[0008] In this configuration, notification data can be automatically provided to the user by triggering whether the result of the first analysis of the target data satisfies the notification condition. In other words, notification data can be provided data-driven, rather than user-triggered. Therefore, with this configuration, it is expected that the effort required for the user to monitor the data to be analyzed (target data) can be reduced.

[0009] In the information processing device according to the above aspect, the notification condition may include a change in the trend of the target data as a result of the first analysis. With this configuration, it is possible to provide notification data at a timing when a change in the trend of the target data occurs.

[0010] In the information processing device according to the above aspect, the first analysis may include calculating statistics of the target data. Trend changes may be evaluated based on the calculated statistics. Trend changes in the target data can be appropriately evaluated based on the statistics. Therefore, with this configuration, notification data can be expected to be provided at an appropriate time.

[0011] In the information processing device according to the above aspect, the target data may include numerical data for each of a plurality of items. The first analysis may include calculating a ranking for each of the plurality of items from the numerical data. The change in trend may be evaluated according to the degree of change in the calculated ranking. The change in ranking may be used to appropriately evaluate the change in trend in the target data. Therefore, with this configuration, notification data can be expected to be provided at an appropriate time.

[0012] In the information processing device according to the above aspect, at least a part of the notification conditions may be generated by the first large-scale generative model from reference data related to the notification conditions. With this configuration, it is expected that preparation of the notification conditions will be simplified.

[0013] In the information processing device according to the above aspect, the reference data may include condition specification data input by a user. With this configuration, it is expected that the effort required by the user to set notification conditions can be reduced.

[0014] In the information processing device according to the above aspect, the reference data may include metadata describing the target data. According to this configuration, the reference data including the metadata can inform the first large-scale generative model of the content of the target data. This allows for the generation of appropriate notification conditions.

[0015] In the information processing device according to the above aspect, the reference data may include at least a portion of the target data. According to this configuration, the reference data includes at least a portion of the target data, so that the content and samples of the target data can be taught to the first large-scale generative model. This allows for the generation of appropriate notification conditions.

[0016] In the information processing device according to the above aspect, the reference data may include a sample of a notification condition. With this configuration, the sample of the notification condition can be taught to the first large-scale generative model. This makes it possible to expect that an appropriate notification condition will be generated.

[0017] In the information processing device according to the above aspect, the reference data may include media data on social trends related to the target data. With this configuration, it is expected that appropriate notification conditions will be generated in accordance with the social trends.

[0018] In the information processing device according to the above aspect, monitoring whether the result of the first analysis satisfies the notification condition may include causing the second large-scale generative model to perform the first analysis by providing at least a portion of the target data as input. This configuration makes it possible to at least partially omit the construction of the analysis process. This is expected to simplify the construction of the system.

[0019] In the information processing device according to the above aspect, monitoring whether the result of the first analysis satisfies the notification condition includes: providing at least a part of the target data as an input to perform a second large-scale generation; The method may include having the synthesis model perform a first analysis and determine whether the result of the first analysis satisfies the notification condition. This configuration makes it possible to at least partially omit the construction of the analysis process and the process of determining whether the notification condition is satisfied. This is expected to simplify the construction of the system.

[0020] In the information processing device according to the above aspect, determining whether the result of the first analysis satisfies the notification condition may be configured by having the second large-scale generative model determine whether the result of the first analysis satisfies the notification condition by referring to media data on social trends related to the target data. With this configuration, it is possible to expect appropriate determination of whether the notification condition is satisfied in accordance with social trends.

[0021] In the information processing device according to the above aspect, monitoring whether the result of the first analysis satisfies the notification condition may include determining whether the result of the first analysis satisfies the notification condition by analyzing explanatory data generated from the result of the first analysis, the explanatory data explaining the result of the first analysis. With this configuration, it can be expected that the explanatory data will enable flexible determination according to the context.

[0022] In the information processing device according to the above aspect, the notification data may include search information regarding search results obtained by providing a search engine with words included in at least one of the results of the second analysis, the notification conditions, and the explanation data. The explanation data may be generated from the results of the second analysis and configured to explain the results of the second analysis. This configuration can be expected to provide notification data with a wide range of content.

[0023] In the information processing device according to the above aspect, the control unit may be further configured to monitor whether a result of a third analysis of the continuously collected target data satisfies a task execution condition, and to perform a predetermined task in response to the result of the monitoring that the result of the third analysis satisfies the task execution condition. With this configuration, the predetermined task can be automatically performed in a data-driven manner. This can be expected to reduce the effort required for a user to monitor the data to be analyzed (target data) when performing the predetermined task.

[0024] In one example, the predetermined task may include any information processing other than notification of notification data. In another example, the predetermined task may include notification of notification data. As a result, the information processing device according to each of the above aspects may be expanded from data-driven notification of notification data to data-driven performance of a predetermined task. For example, an information processing device according to one aspect of the present disclosure may include a control unit. The control unit may be configured to monitor whether a result of analysis of continuously collected target data satisfies a task execution condition, and to perform a predetermined task in response to the result of the monitoring that the analysis result satisfies the task execution condition. This configuration is expected to reduce the effort required by a user to monitor data to be analyzed (target data) when performing a predetermined task.

[0025] Note that the embodiments of the present disclosure may not be limited to the above-described information processing device. As another aspect of the information processing device according to each of the above aspects, one aspect of the present disclosure may be an information processing method that realizes all or part of each of the above configurations, a program, or a machine-readable storage medium storing such a program. Here, the machine-readable storage medium may be a non-transitory medium that stores information such as a program by electrical, magnetic, optical, mechanical, or chemical action. The non-transitory storage medium may include a storage medium (e.g., a CD, a DVD, a semiconductor memory), an auxiliary storage device of a computer, an external storage device connected to a computer, etc.

[0026] For example, an information processing method according to an aspect of the present disclosure may be executed by a computer, and may include monitoring whether a result of a first analysis of continuously collected target data satisfies a notification condition, and, if the result of the first analysis satisfies the notification condition as a result of the monitoring, notifying a user terminal of notification data indicating a result of a second analysis of the target data.

[0027] Furthermore, for example, a program according to an aspect of the present disclosure may be a program for causing a computer to execute an information processing method, which may include monitoring whether a result of a first analysis of continuously collected target data satisfies a notification condition, and, if the result of the first analysis satisfies the notification condition as a result of the monitoring, notifying a user terminal of notification data indicating a result of a second analysis of the target data.

[0028] Furthermore, for example, an information processing method according to an aspect of the present disclosure may be executed by a computer, and may include monitoring whether a result of an analysis of continuously collected target data satisfies a task execution condition, and performing a predetermined task in response to the result of the monitoring that the result of the analysis satisfies the task execution condition.

[0029] Furthermore, for example, a program according to an aspect of the present disclosure may be a program for causing a computer to execute an information processing method, which may include monitoring whether a result of an analysis of continuously collected target data satisfies a task execution condition, and performing a predetermined task in response to the result of the monitoring that the result of the analysis satisfies the task execution condition. [Effects of the Invention]

[0030] According to one aspect of the present disclosure, it is possible to provide a technique for reducing the effort required by a user to monitor data to be analyzed. [Brief explanation of the drawings]

[0031] [Figure 1] FIG. 1 schematically illustrates an example of a situation to which the present disclosure is applied. [Figure 2] FIG. 2 shows an example of a method for setting notification conditions. [Figure 3] FIG. 3 shows a schematic diagram of an example of a method for generating analysis results. [Figure 4] FIG. 4 shows an example of a method for obtaining the analysis results and determining the notification conditions. [Figure 5] FIG. 5 schematically illustrates an example of a method for indirectly determining whether a notification condition is satisfied. [Figure 6] FIG. 6 shows an example of a method for generating notification data. [Figure 7] FIG. 7 is a diagram illustrating an example of a method for notifying notification data. [Figure 8] FIG. 8 is a diagram illustrating an example of a situation in which a response to notification data is received. [Figure 9] FIG. 9 shows a schematic diagram of an example of a scene in which a task is performed. [Figure 10] FIG. 10 is a diagram illustrating an example of a hardware configuration of an information processing device. [Figure 11] FIG. 11 is a diagram illustrating an example of the software configuration of the information processing device. [Figure 12] FIG. 12 is a flowchart illustrating an example of a processing procedure for setting a notification condition by the information processing device. [Figure 13] FIG. 13 is a flowchart illustrating an example of a processing procedure for notifying an analysis result by an information processing device. [Figure 14] FIG. 14 is a flowchart illustrating an example of a processing procedure for task execution by an information processing device. [Figure 15] Figure 15 shows the prompts given to the large-scale generative model to generate pseudo-data. [Figure 16] FIG. 16 shows the generated pseudo data. [Figure 17] FIG. 17 shows the prompts given in Experimental Example 1-1. [Figure 18] FIG. 18 shows the results obtained in Experimental Example 1-1. [Figure 19] FIG. 19 shows the prompts given in Experimental Example 1-2. [Figure 20] FIG. 20 shows the results obtained in Experimental Example 1-2. [Figure 21] FIG. 21 shows the prompts given in Experimental Examples 1-3. [Figure 22] FIG. 22 shows the results obtained in Experimental Example 1-3. [Figure 23] FIG. 23 shows the results obtained in Experimental Example 1-3. [Figure 24] FIG. 24 shows the prompts given in Experimental Examples 1-4. [Figure 25] FIG. 25 shows the results obtained in Experimental Example 1-4. [Figure 26] FIG. 26 shows the prompts given in Experimental Examples 1-5. [Figure 27] FIG. 27 shows the results obtained in Experimental Example 1-5. [Figure 28] FIG. 28 shows the prompts given in Experimental Example 2-1. [Figure 29] FIG. 29 shows the results obtained in Experimental Example 2-1. [Figure 30] FIG. 30 shows the prompts given in Experimental Example 2-2. [Figure 31] FIG. 31 shows the results obtained in Experimental Example 2-2. DETAILED DESCRIPTION OF THE INVENTION

[0032] An embodiment according to one aspect of the present disclosure will be described below with reference to the drawings. However, the embodiment described below is merely an example of the present disclosure in all respects. Various improvements or modifications may be made without departing from the scope of the present disclosure. In implementing the present disclosure, specific configurations according to the embodiment may be appropriately adopted. Note that while data appearing in the present embodiment is described in natural language, more specifically, it may be specified using pseudo-language, commands, parameters, machine language, electrical signals, etc. that can be recognized by a machine such as a computer.

[0033] §1 Application Examples 1 is a diagram illustrating an example of a situation to which the present disclosure is applied. An information processing device 1 according to this embodiment is one or more computers configured to control notification of notification data 5 in a data-driven manner.

[0034] In this embodiment, the information processing device 1 monitors whether or not the first analysis result 30 of the target data 2, which is continuously collected, satisfies the notification condition 4. Monitoring whether or not the first analysis result 30 satisfies the notification condition 4 may include determining whether or not the first analysis result 30 satisfies the notification condition 4. Monitoring (determining) whether or not the first analysis result 30 satisfies the notification condition 4 may be continuously performed.

[0035] If the information processing device 1 determines, as a result of monitoring, that the first analysis result 30 satisfies the notification condition 4, it notifies the user terminal U0 of the notification data 5 indicating the second analysis result 35 of the target data 2. That is, the information processing device 1 executes transmission of the notification data 5 to the user terminal U0 when the first analysis result 30 satisfies the notification condition 4.

[0036] In this embodiment, the notification data 5 can be automatically provided to the user Z0 by using the result 30 of the first analysis of the target data 2 as a trigger when the result 30 satisfies the notification condition 4. In other words, the notification data 5 can be provided in a data-driven manner, rather than being user-triggered. Therefore, according to this embodiment, it is expected that the effort required for the user Z0 to monitor the target data 2 can be reduced.

[0037] [Target data] The target data 2 may be any data that can be the subject of analysis. As long as it can be the subject of analysis, the type and format of the target data 2 are not particularly limited and may be selected appropriately depending on the embodiment. In one example, the target data 2 may include purchase record data such as POS (Point Of Sales) data. In one example, the target data 2 may be data obtained by a sensor. The sensor data 2 may include sensing data obtained by measuring the status of a target house (inside, around, etc.). The type of sensor may be selected arbitrarily. In one example, the sensor may include a vital sensor. The vital sensor may include, for example, a blood pressure monitor, a pulse rate monitor, a heart rate monitor, an electrocardiograph, an electromyograph, a thermometer, a skin electrodermal response monitor, a microwave sensor, an electroencephalograph, a magnetoencephalograph, an activity monitor, a blood glucose monitor, an electrooculography sensor, an eye movement monitor, etc. In one example, the sensor may include an IoT (Internet of Things) sensor. The IoT sensor may include, for example, an image sensor, an infrared sensor, an activity (audio) sensor, a gas (CO2, etc.) sensor, a current sensor, a smart meter (a sensor that measures the power consumption of home appliances, lighting, etc.), etc. The IoT sensor may be used to observe the status of the target house (inside, around, etc.). In one example, the target data 2 may include financial data such as stock price data. In another example, the target data 2 may include payment data such as a credit payment history.

[0038] The target data 2 may be generated as appropriate. The target data 2 may be generated in response to any activity, such as purchasing behavior, sensing, financial transactions, or payment processing. The data source DS of the target data 2 is not particularly limited and may be selected as appropriate depending on the embodiment. In one example, the data source DS may include a computer, and the target data 2 may be generated by information processing by the computer. The computer may include a server device, a user terminal, etc. In one example, the data source DS may include a sensor, and the target data 2 may be generated by processing by the sensor. The target data 2 may be generated directly by the sensor, or may be generated by information processing by the computer of raw data obtained by the sensor.

[0039] The target data 2 collected from the data source DS may be appropriately stored in a memory resource MR. The configuration of the memory resource MR (storage device) is not particularly limited and may be appropriately selected depending on the embodiment. In one example, at least a portion of the memory resource MR may be disposed within the information processing device 1. That is, the memory resource of the information processing device 1 may be used as at least a portion of the memory resource MR that stores the target data 2. In one example, at least a portion of the memory resource MR may be disposed outside the information processing device 1. That is, at least a portion of the memory resource MR may be configured as an external storage device other than the information processing device 1. At least a portion of the target data 2 may be subjected to analysis processing immediately, or may be subjected to analysis processing after being stored for an arbitrary period of time.

[0040] [analysis] The analysis (first analysis, second analysis) may be configured by any arithmetic processing of the target data 2. The method of the analysis (first analysis, second analysis) is not particularly limited and may be appropriately selected depending on the embodiment. In one example, the analysis (first analysis, second analysis) may include statistical analysis. The statistical analysis may include, for example, ranking, calculation of statistics, detection of outliers, etc. The statistics may include, for example, the mean, variance, standard deviation, median, maximum value, minimum value, Nth percentile value, sum value, cumulative value, etc. The arithmetic processing of the analysis (first analysis, second analysis) may be defined by any mathematical formula. A trained machine learning model may be used in the analysis (first analysis, second analysis) (such as the cases in Figures 3 and 4 described below). The trained machine learning model may include a large-scale generative model such as a large-scale language model (LLM), a large-scale visual language model (VLM), or a large-scale speech model. A known method may be adopted as the analysis method.

[0041] The analysis (first analysis, second analysis) may be performed for any purpose. In one example, if the target data 2 includes purchase record data, the analysis (first analysis, second analysis) may include an analysis of purchasing trends. The analysis of purchasing trends may include, for example, calculating the ranking of product sales. In one example, if the target data 2 includes sensing data, the analysis (first analysis, second analysis) may include an analysis of trends related to the object observed by the sensor. The analysis of trends related to the object observed may include, for example, whether vital signs satisfy a predetermined condition (e.g., the measured blood glucose level is in a hypoglycemic state), whether the home situation satisfies a predetermined condition (e.g., no one is at home), etc. In one example, if the target data 2 includes financial data, the analysis (first analysis, second analysis) may include an analysis of trends related to financial The analysis of financial trends may include, for example, calculating fluctuations in stock prices and determining whether the calculated fluctuations satisfy predetermined conditions (such as stop-loss criteria). In one example, when the target data 2 includes payment data, the analysis (first analysis, second analysis) may include an analysis of payment trends. The analysis of settlement trends may include, for example, detecting whether at least one of the amount and the purchased item deviates from the purchase history (an example of outlier detection).

[0042] The range of the target data 2 to be processed in the analysis (first analysis, second analysis) may be determined appropriately depending on the embodiment. In one example, the entire collected target data 2 may be the target of the analysis. In another example, a portion of the target data 2 (for example, data within a predetermined range from the current time) may be the target of the analysis.

[0043] (Relationship between the first and second analyses) The first analysis is performed to determine whether the notification condition 4 is satisfied. The second analysis is performed to generate notification data 5 to be provided to the user Z0. In one example, the first analysis may be the same as (common to) the second analysis. In this case, the result of one analysis process may be used as the first analysis result 30 and the second analysis result 35. In one example, the first analysis may be included in the second analysis as part of the second analysis. In this case, the first analysis may be performed as part of the second analysis. The notification data 5 may be transmitted to the user terminal U0 when a part of the analysis result (the second analysis result 35) constituting the notification data 5 satisfies the notification data 5. In one example, the second analysis may be included in the first analysis as part of the first analysis. In this case, the second analysis may be performed as part of the first analysis. The notification data 5 may be composed of a part of the analysis result (the first analysis result 30) that is the target of determining whether the notification condition 4 is satisfied. In one example, the first analysis and the second analysis may partially coincide. In this case, a part of the analysis may be performed commonly as the first analysis and the second analysis. Some of the analysis results may be used as part of the first analysis results 30 and part of the second analysis results 35. In one example, the first analysis may be different from the second analysis. In this case, the first and second analyses may be performed separately. The order of processing may be arbitrary.

[0044] (First and second analysis execution timing) The first analysis and the second analysis may be performed at any timing. In one example, it may be determined whether a predetermined condition is satisfied, and the first analysis may be performed in response to the satisfaction of the predetermined condition. The predetermined condition may include a time condition. The time condition may include, for example, the arrival of a specific time (date, day of the week, time of day, etc.), the passage of a predetermined period, etc. In this case, the first analysis may be performed in response to the satisfaction of the time condition (each time a specific time arrives, each time a predetermined period occurs, etc.). Furthermore, the predetermined condition may include a data volume condition. The data volume condition may include, for example, the addition of an amount of data exceeding a threshold, the collection of unknown data (such as purchase record data including a new product ID), etc. In this case, the first analysis may be performed in response to the satisfaction of the data volume condition (each time an amount of data exceeding a threshold is added, each time unknown data is added, etc.).

[0045] In one example, the second analysis may include analysis items common to the first analysis (such as when the first analysis and the second analysis are the same). In this case, the analysis of the items (second analysis) may be performed in common with the first analysis. The second analysis may also include analysis items different from those in the first analysis (such as when the first analysis and the second analysis are different from each other). In this case, the analysis of the items different from those in the first analysis (second analysis) may be performed before the first analysis, together with the first analysis, or after the first analysis. If the analysis is performed after the first analysis, the analysis of the items different from those in the first analysis (second analysis) may be performed only if the result 30 of the first analysis satisfies the notification condition 4, or may be performed regardless of whether the result 30 of the first analysis satisfies the notification condition 4.

[0046] (Monitoring process execution timing) Furthermore, monitoring of whether or not the result 30 of the first analysis satisfies the notification condition 4 (processing for determining whether or not the notification condition 4 is satisfied) may be performed at any timing. The process for determining whether or not the result 30 of the first analysis is satisfied may be performed at any timing after the result 30 of the first analysis is obtained. In one example, the process for determining whether or not the notification condition 4 is satisfied may be performed in response to the execution of the first analysis and the acquisition of the result 30 of the first analysis. In other words, the process for determining whether or not the notification condition 4 is satisfied may be performed as a series of processes following the execution of the first analysis.

[0047] [Notification conditions] The notification condition 4 is a trigger condition for determining whether or not to execute notification of the notification data 5. The notification condition 4 is not particularly limited, and may be defined appropriately depending on the embodiment, such as the purpose of notifying the notification data 5.

[0048] In one example, the notification condition 4 may include a change in the trend of the target data 2 in the first analysis result 30. That is, the notification condition 4 may include a condition (criteria, etc.) defined so as to evaluate sufficiency according to a change in the trend of the target data 2. The notification condition 4 may be defined appropriately to capture a change in the trend of the target data 2. According to one example of the present embodiment, notification data 5 can be provided to user Z0 when a change in the trend of the target data 2 occurs. Furthermore, the main purpose of data analysis is often to investigate whether a change in the trend of the data has occurred. Therefore, according to one example of the present embodiment, it is possible to achieve the main purpose of data analysis while reducing the effort required to monitor the data (target data 2). Note that a change in the trend of the target data 2 may be evaluated using any method. According to one example of the present embodiment, it may be determined whether a change in the trend has occurred (i.e., whether the notification condition 4 is satisfied) using at least one of the following two methods.

[0049] (First evaluation method) In one example, the first analysis may include calculating a statistic 301 of the target data 2. A change in trend may be evaluated according to the calculated statistic 301. The statistic 301 may include, for example, a mean value, a variance, a standard deviation, a median, a maximum value, a minimum value, an Nth percentile value, a sum value, a cumulative value, etc.

[0050] The method for evaluating a trend change using the statistic 301 is not particularly limited and may be appropriately defined depending on the embodiment. In one example, whether a trend change has occurred (i.e., whether notification condition 4 is satisfied) may be determined by comparing the statistic 301 with a threshold. For example, if the target data 2 includes purchase record data, the calculated statistic 301 may include the total sales value of the target products. In this case, if the total sales value of the target products exceeds the threshold, it may be determined that notification condition 4 is satisfied (a trend change has occurred), and notification data 5 may be transmitted to the user terminal U0. In one example, an outlier may be detected in the target data 2 using the statistic 301 (such as a mean value, variance, or standard deviation). A known method may be used to detect the outlier. Whether a trend change has occurred may be determined based on the number of detected outliers. The threshold for the number of occurrences used to determine whether a trend change has occurred may be defined arbitrarily. Furthermore, in one example, evaluating a trend change based on the statistic 301 may be performed by evaluating a trend change based on a fluctuation (amount of change) in the statistic 301. For example, if the target data 2 includes purchase record data, the calculated statistics 301 may include the average age (average age) of the purchasing users. In this case, if the change in the average age exceeds a threshold, it may be determined that the notification condition 4 is satisfied (a trend change has occurred), and notification data 5 may be sent to the user terminal U0.

[0051] The statistics 301 can appropriately evaluate a change in trend in the target data 2. Therefore, according to an example of this embodiment, it is possible to expect that the notification data 5 will be provided at an appropriate timing when a change in trend occurs in the target data 2.

[0052] (Second evaluation method) Also, in one example, the target data 2 may include numerical data 20 relating to each of a plurality of items. The items may be any matter (things, people, events, etc.) that can be ranked. For example, if the target data 2 includes purchase record data, the items may be products, product categories, vendors, purchasing users, etc. The numerical data 20 may be composed of one or more arbitrary numerical values ​​that can be ranked, such as price, quantity, age, measurement value, stock price, payment amount, etc.

[0053] The first analysis may include calculating a rank 303 of each of the multiple items from the numerical data 20. The rank 303 may be derived directly from the numerical data 20, or may be derived indirectly from the numerical data 20 through calculation of statistics (average, variance, standard deviation, median, maximum, minimum, Nth percentile value, sum, cumulative value, etc.). For example, if the target data 2 includes purchase record data and the items are products, the rank 303 may be calculated from at least one of the product sales, sales quantity, and the amount of change therein. The amount of change may be constituted by the degree of increase or decrease in at least one of the sales and sales quantity.

[0054] A change in trend may be evaluated according to the degree of change in the calculated rank 303. A threshold for the degree of change for determining that a change in trend has occurred may be defined arbitrarily. In one example, when a change in rank 303 has occurred in a number of items exceeding (or equal to or greater than) the threshold, it may be determined that notification condition 4 is satisfied (a change in trend has occurred), and notification data 5 may be sent to the user terminal U0. In one example, when a change in rank 303 has occurred (threshold 0), it may be determined that notification condition 4 is satisfied, and notification data 5 may be sent to the user terminal U0. For example, in the case of the above-mentioned purchase record data, when a change has occurred in the ranking of product sales or sales volume, it may be determined that notification condition 4 is satisfied, and notification data 5 may be sent to the user terminal U0.

[0055] The change in the ranking 303 makes it possible to appropriately evaluate the change in the trend in the target data 2. Therefore, according to an example of this embodiment, it is possible to expect that the notification data 5 will be provided at an appropriate timing when a change in the trend in the target data 2 occurs.

[0056] (others) In one example, the first analysis may include calculating a statistic 301 of the target data 2 and a rank 303 of each item. A change in trend may be evaluated according to at least one of the degree of change in the calculated statistic 301 and the rank 303. A change in trend may also be evaluated according to both the degree of change in the calculated statistic 301 and the rank 303.

[0057] When evaluating whether a change in trend has occurred based on the fluctuation (amount of change) of at least one of the statistic 301 and the rank 303, the fluctuation (amount of change) of the statistic 301 and the rank 303 may be calculated arbitrarily. In one example, at least one of the statistic 301 and the rank 303 may be calculated every predetermined period, and the fluctuation (amount of change) may be calculated between two adjacent periods (for example, the previous calculation result and the current calculation result). The predetermined period may be defined as appropriate depending on the embodiment.

[0058] [Notification condition data format / setting method] As long as it is possible to determine whether or not the notification condition 4 is satisfied, the data format of the notification condition 4 is not particularly limited and may be selected appropriately depending on the embodiment. In one example, at least a part of the notification condition 4 may be defined in a numerical format such as the threshold value. For example, in a situation where a change in trend is evaluated using the total sales value of the target product, the notification condition 4 may be defined by a threshold value for the total sales value. In one example, at least a part of the notification condition 4 may be defined in a data format other than numerical values, such as text (such as the case of FIG. 4 described later). There is no particular limit to the number of conditions constituting the notification condition 4. It is not necessary to set the number of the spools, and it may be determined appropriately depending on the embodiment.

[0059] Furthermore, the method for setting the notification conditions 4 is not particularly limited and may be appropriately selected depending on the embodiment. In one example, at least a portion of the notification conditions 4 may be provided in advance. At least a portion of the notification conditions 4 may be predefined in a program (such as the program 81 described below). In one example, at least a portion of the notification conditions 4 may be manually set by an operator. For example, the information processing device 1 may receive a condition setting operation from an operator directly or indirectly via a terminal, and may set the notification conditions 4 in accordance with the received setting operation. The setting operation may include inputting, specifying (selecting), etc. of conditions. The specification of the notification conditions 4 by the operator (user) may include, for example, selecting parameters (such as the statistic 301 and the rank 303) and specifying parameter values ​​(such as thresholds). In one example, at least a portion of the notification conditions 4 may be automatically set at least in part by information processing, such as using a large-scale generative model.

[0060] FIG. 2 schematically shows an example of a method for setting notification conditions 4 according to this embodiment. In one example, at least a part of the notification conditions 4 may be generated by a first large-scale generative model M1 from reference data 40 related to the notification conditions 4. The first large-scale generative model M1 is a large-scale generative model used to generate the notification conditions 4. The type of the first large-scale generative model M1 is not particularly limited and may be selected appropriately depending on the embodiment. The first large-scale generative model M1 may be a trained model generated by machine learning using a large amount of training data. The first large-scale generative model M1 may be, for example, a large-scale language model, a large-scale visual language model, a large-scale speech model, etc. The first large-scale generative model M1 may include a multimodal model. The configuration and number of parameters of the first large-scale generative model M1 may be determined appropriately depending on the embodiment. The first large-scale generative model M1 may be, for example, a Transformer, a diffusion model, etc. The first large-scale generative model M1 may have any structure such as a neural memory, a persistent memory, or the like. The first large-scale generative model M1 may have any mechanism such as an attention mechanism, a memory (Neural Memory, Persistent Memory, etc.). The first large-scale generative model M1 may be fine-tuned depending on the purpose of use. As a specific example, the first large-scale generative model M1 may be a model based on Claude, GPT, CyberAgentLM, etc. A known model of may be used.

[0061] The computational processing of the first large-scale generative model M1 may be executed on any computer. In one example, the first large-scale generative model M1 may be deployed on the information processing device 1. The generation process of the notification condition 4 by the first large-scale generative model M1 (the process of generating at least a part of the notification condition 4) may be executed on the information processing device 1. In another example, the first large-scale generative model M1 may be deployed on a computer other than the information processing device 1. The generation process of the notification condition 4 by the first large-scale generative model M1 may be executed on the other computer. The information processing device 1 may obtain at least a part of the notification condition 4 by receiving the computational results of this generation process from the other computer.

[0062] The instruction given to the first large-scale generative model M1 when generating the notification condition 4 may be configured appropriately depending on the embodiment. In one example, a prompt P1 including the reference data 40 and the generation instruction I1 may be given to the first large-scale generative model M1 to cause the first large-scale generative model M1 to generate the notification condition 4. As long as it is possible to instruct the generation of the notification condition 4, the configuration of the generation instruction I1 is not particularly limited and may be determined appropriately depending on the embodiment. In one example, the generation instruction I1 may be given in natural language, such as "Please generate conditions for notifying a report (notification data 5) taking into consideration AAA (reference data 40)." The data format of the generation instruction I1 is not particularly limited and may be selected appropriately depending on the embodiment.

[0063] The method of making the first large-scale generative model M1 refer to the reference data 40 is not limited to this example, and may be changed as appropriate depending on the embodiment. At least a portion of the reference data 40 may be acquired during the calculation process of the first large-scale generative model M1, rather than being provided as a prompt P1. For example, the generation instruction I1 may include an instruction to refer to the reference data 40. In response to this, the first large-scale generative model M1 may acquire at least a portion of the reference data 40 during the process of generating the notification condition 4, and may use the acquired reference data 40 to generate the notification condition 4. The reference data 40 may be acquired by any method. For example, the reference data 40 may be stored in a predetermined storage device (such as a memory resource MR). A computer executing the calculation process of the first large-scale generative model M1 may acquire at least a portion of the reference data 40 by accessing a predetermined storage area. Also, for example, the instruction to refer to the reference data 40 may include an instruction to search for the reference data 40 externally. In response to this, at least a portion of the reference data 40 may be acquired by searching the Internet using a search engine. A known search engine may be used as the search engine. Search conditions may be specified arbitrarily.

[0064] The generation result obtained by the first large-scale generative model M1 may be appropriately reflected in the setting of the notification condition 4. In one example, the generation result of the first large-scale generative model M1 may be used as the notification condition 4 as is. In one example, an operator may reflect the generation result of the first large-scale generative model M1 in the setting of the notification condition 4 by operating the information processing device 1 directly or indirectly via a terminal. The operator's operation may include approval, correction, etc. The operator may reflect the generation result of the first large-scale generative model M1 as is in the setting of the notification condition 4, or may modify the generation result before reflecting it in the setting of the notification condition 4. In one example, the generation result of the first large-scale generative model M1 may be reflected in the setting of the notification condition 4 after undergoing any calculation processing (such as rule-based conversion).

[0065] For example, when the notification condition 4 is defined in a numerical format such as a threshold, the first large-scale generative model M1 may be caused to propose a numerical value to be set as the notification condition 4. The numerical condition (generation result) proposed by the first large-scale generative model M1 may be set as the notification condition 4. Furthermore, for example, the generation instruction I1 may include an instruction to generate program code for determining the notification condition 4. In response to this, the generation result of the first large-scale generative model M1 may include program code that can be used to determine the notification condition 4. The information processing device 1 may set the notification condition 4 by incorporating the generated program code into the monitoring (determination) process, and may also reflect the setting of the notification condition 4 in the monitoring (determination) process.

[0066] In one example of this embodiment, by using the first large-scale generative model M1, at least a part of the process of setting the notification conditions 4 can be automated. By utilizing the knowledge acquired by the first large-scale generative model M1, appropriate notification conditions 4 can be obtained even if the user is not familiar with the items of the target data 2. Therefore, according to this example of the present embodiment, it is expected that the preparation of the notification conditions 4 can be simplified. Furthermore, when the first large-scale generative model M1 is configured as a multimodal model, the notification conditions 4 can be generated using data in any format, such as text, audio, or image. The notification conditions 4 may be configured to include, as a result generated by the first large-scale generative model M1, that a change in the trend of the target data 2 has occurred in the first analysis result 30.

[0067] The configuration of the reference data 40 is not particularly limited as long as it is related to the notification condition 4. The reference data 40 may include any information that can be used to generate the notification condition 4.

[0068] (Conditional data) In one example, the reference data 40 may include condition specification data 401 input by a user Z1. The user Z1 may be the same as or different from the user Z0 to whom the notification data 5 is to be sent. The user terminal U1 is used to input the condition specification data 401. The user terminal U1 may be the same as or different from the user terminal U0 to which the notification data 5 is to be notified.

[0069] In one example, the condition designation data 401 may be obtained by user Z1 freely inputting the desired notification condition 4. The data format of the condition designation data 401 is not particularly limited and may be selected appropriately depending on the embodiment. In one example, the condition designation data 401 may be composed of at least one of text data, audio data, and image data. The condition designation data 401 may include reference data designated by user Z1. The reference data may be any data that can be used to generate the notification condition 4. In other words, being input by user Z1 may include being designated by user Z1.

[0070] Also, in one example, the condition designation data 401 may be configured to directly or indirectly designate notification condition 4. For example, when the target data 2 includes purchase record data, the condition designation data 401 may include a direct description of notification condition 4, such as "Please notify me if BBB's sales fluctuate by C%." Also, for example, the condition designation data 401 may include an indirect description of notification condition 4, such as "Please notify me if there are any interesting changes" or "Please notify me if there are signs of a change in customer purchasing trends."

[0071] According to one example of this embodiment, since the reference data 40 includes the condition specification data 401, a user (user Z1) can obtain appropriate notification conditions 4 even if he or she is not familiar with the items of the target data 2. For example, when evaluating a trend change using a numerical value (such as a threshold), the notification conditions 4 can be obtained by inputting text or the like without directly specifying the reference numerical value. Therefore, it is expected that the user's effort in setting the notification conditions 4 can be reduced.

[0072] (Metadata) In one example, the reference data 40 may include metadata 403 that describes the target data 2. The metadata 403 may be configured to include descriptions that describe the contents of the fields (items) of the target data 2, domain information, etc. The domain information may include, for example, acquisition conditions, environmental conditions (such as conditions of the sensor measurement environment), etc. For example, if the target data 2 includes purchase record data, the acquisition conditions or environmental conditions may include the store situation, number of customers, customer demographics, customer type, weather, temperature, humidity, etc.

[0073] The metadata 403 may be obtained in any manner. In one example, the metadata 403 may be provided manually by an operator. In one example, at least a portion of the metadata 403 may be automatically generated from the subject data 2. For example, the metadata 403 may be automatically generated from the subject data 2 using a trained machine learning model, such as a large-scale generative model.

[0074] According to an example of this embodiment, the reference data 40 includes the metadata 403, so that the contents of the target data 2 can be taught to the first large-scale generative model M1. As a result, it can be expected that an appropriate notification condition 4 will be generated.

[0075] (At least part of the data) In one example, the reference data 40 may include at least a portion of the target data 2. The amount of target data 2 provided to the first large-scale generative model M1 as reference data 40 is not particularly limited and may be determined appropriately depending on the embodiment. The range of target data 2 provided as reference data 40 may or may not overlap with the target range of the first analysis or the second analysis. According to one example of this embodiment, by having the reference data 40 include at least a portion of the target data 2, the contents and samples of the target data 2 can be taught to the first large-scale generative model M1. This can be expected to generate appropriate notification conditions 4.

[0076] (sample) In one example, the reference data 40 may include samples 405 of notification conditions 4. The samples 405 may be configured to exemplify at least some of the notification conditions 4. For example, at least some of the notification conditions 4 that have been used in the past may be used as the samples 405. At least some of the notification conditions 4 that have been generated in the past by the first large-scale generative model M1 may be used as the samples 405. According to one example of the present embodiment, by teaching the samples 405 of the notification conditions 4 to the first large-scale generative model M1, it is possible to expect that appropriate notification conditions 4 will be generated.

[0077] (Media Data) In one example, the reference data 40 may include media data 407 related to social trends related to the target data 2. The media may include, for example, mass media, digital media, etc. The mass media may include, for example, newspapers, television, magazines, etc. The magazines may include trade papers, industry news, academic journals, newsletters, etc. The digital media may include, for example, news sites, social media (SNS: Social Networking Service, etc.), video platforms, blogs, curation sites, etc. Digital media may include applications such as the news sites mentioned above.

[0078] Social trends may include any situation related to target data 2 and communicated through the media. For example, if target data 2 includes purchase record data, media data 407 may include any information related to at least one of the products, stores, and customers for which purchase record data is collected, communicated through the media. As a specific example, if the products for which purchase record data is collected include beer, media data 407 may include news about beer sales (e.g., news reporting that beer sales are on the rise). If the products for which purchase record data is collected are products sold at convenience stores, media data 407 may include news about convenience stores. Media data 407 may include news about the area to which the store selling the products belongs. For example, if the store is located in Kyoto, media data 407 may include news reporting events that occurred in Kyoto.

[0079] As long as the data format of the media data 407 can be input to the first large-scale generative model M1, it is not particularly limited and may be selected appropriately depending on the embodiment. The media data 407 may be appropriately provided to the first large-scale generative model M1 as reference data 40. In one example, the media data 407 may be collected in advance. At least a portion of the collected media data 407 may be provided to the first large-scale generative model M1 as a prompt P1 or during the calculation process. Furthermore, the generation instruction I1 (an instruction to refer to the reference data 40) may include an instruction to search for the media data 407, so that the media data 407 may be appropriately acquired by having the first large-scale generative model M1 search for it. The instruction to search for the media data 407 is an example of an instruction to search for the reference data 40 from an external source. The information source SO1 of the media data 407 may be selected arbitrarily. The information source SO1 may include a website, a predetermined storage device (memory resource MR, etc.), etc.

[0080] According to one example of this embodiment, the reference data 40 includes media data 407, which allows the first large-scale generative model M1 to learn about social trends communicated in the media. This makes it possible to expect that an appropriate notification condition 4 will be generated in accordance with the social trend. For example, if the media data 407 includes news about beer sales, it is possible to expect that notification condition 4 will be generated that reflects the beer sales trends reported in the news. For example, if the media data 407 includes news about convenience stores, it is possible to expect that notification condition 4 will be generated that reflects the trends in convenience stores reported in the news. In addition, for example, if the media data 407 includes news reporting an event that occurred in Kyoto, it is possible to expect the generation of notification condition 4 that reflects the event that occurred in Kyoto that is being reported in the news.

[0081] In order to reflect social trends close to the time of operation in the notification conditions 4, it is preferable that the media data 407 be composed of information conveying social trends close to the time of generating the notification conditions 4. Therefore, the media data 407 provided to the first large-scale generative model M1 may be narrowed down from all data usable as the media data 407 to data indicating social trends in the same or similar time period as the time of generating the notification conditions 4. In one example, the media data 407 may be composed of information conveyed by media within a predetermined period from the time of generating the notification conditions 4 (e.g., news within k months). The media data 407 may also be composed of information conveyed by media in the same or similar season as the time of generating the notification conditions 4 (e.g., news distributed in the summer of the previous year if the time of generating the notification conditions 4 is summer).

[0082] (others) In one example, the reference data 40 may include at least one of the condition specification data 401, the metadata 403, at least a portion of the target data 2, the sample 405, and the media data 407. In one example, the reference data 40 may be composed of data other than these. The reference data 40 may further include other data in addition to at least one of the condition specification data 401, the metadata 403, at least a portion of the target data 2, the sample 405, and the media data 407. The data used as the reference data 40 may be selected by a user (such as user Z1).

[0083] [Analysis / Decision-Making Entity] The analyses (first analysis, second analysis) may be performed on any computer. In one example, the first analysis and the second analysis may be performed on the information processing device 1. In another example, at least one of the first analysis and the second analysis may be performed on a computer other than the information processing device 1. The computer that performs the first analysis may be at least partially the same as or different from the computer that performs the second analysis.

[0084] Similarly, the process of determining whether or not the first analysis result 30 satisfies the notification condition 4 (the process of determining whether or not the notification condition 4 is satisfied) may also be executed by any computer. In one example, the process of determining whether or not the notification condition 4 is satisfied may be executed on the information processing device 1. In another example, the process of determining whether or not the notification condition 4 is satisfied may be executed on another computer. In this case, determining whether or not the notification condition 4 is satisfied may be configured by obtaining a determination result of whether or not the notification condition 4 is satisfied that is calculated by the other computer. That is, in one example, the information processing device 1 may be configured to monitor whether or not the notification condition 4 is satisfied by receiving the determination result by the other computer.

[0085] The process of determining whether notification condition 4 is satisfied may be defined by any mathematical formula. In one example, the process of determining whether notification condition 4 is satisfied may be rule-based. The rule (notification condition 4) may be defined as appropriate. For example, parameters of the rule (notification condition 4), such as the sales amount, average age, and number of sales of the target product in the case of the above-mentioned statistics 301 and ranking 303, may be pre-defined. When the parameters are pre-defined, setting notification condition 4 may be configured by specifying the parameter value (threshold value, etc.). As described above, the parameter value may be manually specified by an operator or may be at least partially automatically specified by information processing. As described above, in one example, the process of determining whether notification condition 4 is satisfied may be configured by program code generated by the first large-scale generative model M1. Furthermore, a trained machine learning model may be used in the process of determining whether notification condition 4 is satisfied. Training The machine learning models may include large-scale generative models.

[0086] (Using large-scale generative models in analytical processing) FIG. 3 schematically illustrates an example of a method for generating an analysis result according to this embodiment. In one example, monitoring whether the result 30 of the first analysis satisfies the notification condition 4 may include causing the second large-scale generative model M2 to perform the first analysis by providing at least a portion of the target data 2 as input. The second large-scale generative model M2 is a large-scale generative model used to generate the result 30 of the first analysis. Except for the difference in the intended use, the second large-scale generative model M2 may be configured similarly to the first large-scale generative model M1. That is, the type of the second large-scale generative model M2 is not particularly limited and may be selected appropriately depending on the embodiment. The second large-scale generative model M2 may be, for example, a large-scale language model, a large-scale visual language model, a large-scale speech model, etc. The second large-scale generative model M2 may include a multimodal model. The configuration and the number of parameters of the second large-scale generative model M2 may be determined appropriately depending on the embodiment. The second large-scale generative model M2 may be fine-tuned depending on the intended use. As a specific example, the second large-scale generative model M2 may be, for example, a large-scale generative model based on Claude, GPT, etc. , CyberAgentLM, or other known models may be used. In one example, the second large-scale generative model M2 may be the same as the first large-scale generative model M1. That is, one large-scale generative model may be commonly used as the first large-scale generative model M1 and the second large-scale generative model M2. In another example, the second large-scale generative model M2 may be a large-scale generative model different from the first large-scale generative model M1.

[0087] The computational processing of the second large-scale generative model M2 may be executed by any computer. In one example, as shown in FIG. 3, the second large-scale generative model M2 may be deployed in the information processing device 1, and the computational processing of the second large-scale generative model M2 may be executed on the information processing device 1. In another example, the second large-scale generative model M2 may be deployed in a computer other than the information processing device 1, and the computational processing of the second large-scale generative model M2 may be executed on the other computer. The other computer may execute the first analysis processing by the second large-scale generative model M2 at any timing. In one example, the information processing device 1 may issue a first analysis instruction to the other computer, causing the other computer to execute the first analysis processing by the second large-scale generative model M2. In another example, the other computer may autonomously execute the first analysis processing by the second large-scale generative model M2, regardless of an instruction from the information processing device 1. When the process of determining whether the results 30 of the first analysis satisfy the notification condition 4 is executed on the information processing device 1, the information processing device 1 may appropriately obtain the results 30 of the first analysis generated by the second large-scale generative model M2 from another computer.

[0088] The instructions given to the second large-scale generative model M2 when performing the first analysis may be configured appropriately depending on the embodiment. For example, a prompt P20 including at least a portion of the target data 2 and an analysis instruction I20 may be given to the second large-scale generative model M2 to cause the second large-scale generative model M2 to perform the first analysis (i.e., generate the first analysis result 30). As with the reference data 40, at least a portion of the target data 2 may be referenced during the calculation process rather than being given as the prompt P20. The analysis instruction I20 may include analysis conditions such as the type of statistic 301 to be calculated and the period for which the statistic 301 or the rank 303 is to be calculated. As long as the analysis instruction I20 can instruct the performance of the first analysis, the configuration of the analysis instruction I20 is not particularly limited and may be determined appropriately depending on the embodiment. For example, the analysis instruction I20 may be given in natural language, such as "Calculate EEE (statistic 301)," "Calculate the FFF rank," or "Statistically analyze GGG (target data 2)." The data format of the analysis instructions I20 is not particularly limited and may be selected appropriately depending on the embodiment.

[0089] According to one example of this embodiment, a large-scale generative model (second large-scale generative model M2) is used. By doing so, it is possible to at least partially omit the construction of the first analysis process, which is expected to simplify the construction of a system (including the information processing device 1) for executing a series of information processes, such as performing the first analysis, determining whether the notification condition 4 is satisfied, and notifying the notification data 5 when the notification condition 4 is satisfied.

[0090] The first analysis result 30 generated by the second large-scale generative model M2 may be subjected to a process for determining whether the notification condition 4 is satisfied, as appropriate. In one example, following the process for generating the first analysis result 30, the second large-scale generative model M2 may also execute a process for determining whether the first analysis result 30 satisfies the notification condition 4 (such as the form of FIG. 4 described below). In another example, a large-scale generative model other than the second large-scale generative model M2 may be used in the process for determining whether the first analysis result 30 satisfies the notification condition 4. In yet another example, the process for determining whether the first analysis result 30 satisfies the notification condition 4 may be executed by a method other than using a large-scale generative model (such as a rule-based method).

[0091] In one example, the second large-scale generative model M2 may also be used in the process of generating the second analysis result 35 (performing the second analysis). Except for replacing the first analysis with the second analysis, the method of causing the second large-scale generative model M2 to generate the second analysis result 35 may be similar to the method of generating the first analysis result 30 described above. In another example, a large-scale generative model different from the second large-scale generative model M2 may be used in the process of generating the second analysis result 35. In yet another example, the process of generating the second analysis result 35 may be performed by a method other than using a large-scale generative model.

[0092] (Using large-scale generative models in analytical and decision-making processes) FIG. 4 schematically illustrates an example of a method for obtaining analysis results according to the present embodiment and determining whether the notification condition 4 is satisfied. In one example, monitoring whether the first analysis result 30 satisfies the notification condition 4 may include providing at least a portion of the target data 2 as input to a second large-scale generative model M2 to perform the first analysis and determine whether the first analysis result 30 satisfies the notification condition 4. In this example, the second large-scale generative model M2 is a large-scale generative model used to generate the first analysis result 30 and determine whether the notification condition 4 is satisfied. The embodiment of FIG. 4 may be configured similarly to the embodiment of FIG. 3, except that the process for determining whether the notification condition 4 is satisfied is performed by the second large-scale generative model M2. When the computational processing of the second large-scale generative model M2 is executed on another computer, the information processing device 1 may obtain the determination result of whether the notification condition 4 is satisfied from the other computer.

[0093] The instructions given to the second large-scale generative model M2 may be configured appropriately depending on the embodiment. In one example, a prompt P21 including at least a portion of the target data 2, an analysis instruction I21, and a determination instruction I23 may be given to the second large-scale generative model M2 to cause the second large-scale generative model M2 to execute the first analysis process and the process of determining whether the notification condition 4 is satisfied. At least a portion of the target data 2 may be referenced during the calculation process rather than being given as the prompt P21. The analysis instruction I21 may be configured similarly to the above-mentioned analysis instruction I20. The analysis instruction I21 may include analysis conditions.

[0094] Furthermore, as long as it is possible to instruct the execution of the judgment process, the configuration of the judgment instruction I23 is not particularly limited and may be determined appropriately depending on the embodiment. In one example, the judgment instruction I23 may be given in natural language, such as "Please execute notification of the report (notification data 5) when the conditions are met." The data format of the judgment instruction I23 is not particularly limited and may be selected appropriately depending on the embodiment. In order to teach the second large-scale generative model M2 the notification conditions 4, the prompt P21 may include at least one of at least a part of the notification conditions 4 and the condition instruction 4I. In one example, the notification conditions 4 given to the second large-scale generative model M2 may be predefined conditions, or may be conditions generated by the first large-scale generative model M1. The condition instruction 4I may be configured using any information that can be used to cause the second large-scale generative model M2 to derive the notification condition 4. In one example, the condition instruction 4I may be configured similarly to the reference data 40. This may cause the second large-scale generative model M2 to derive the notification condition 4 in the same manner as causing the first large-scale generative model M1 to generate the notification condition 4. Then, the second large-scale generative model M2 may be caused to execute a process of determining whether the derived notification condition 4 is satisfied. At least one of the notification condition 4 and the condition instruction 4I may be incorporated into the determination instruction I23.

[0095] According to one example of this embodiment, by using a large-scale generative model (second large-scale generative model M2), it is possible to at least partially omit the construction of the first analysis process and the process for determining whether the notification condition 4 is satisfied. This is expected to simplify the construction of a system (including the information processing device 1) for executing a series of information processes that perform the first analysis, determine whether the notification condition 4 is satisfied, and notify the notification data 5 when the notification condition 4 is satisfied.

[0096] The information provided to the second large-scale generative model M2 is not limited to the above example and may be modified as appropriate depending on the embodiment. In one example, determining whether the first analysis result 30 satisfies the notification condition 4 may be configured by having the second large-scale generative model M2 determine whether the first analysis result 30 satisfies the notification condition 4 by referring to media data D25 related to social trends related to the target data 2. The media data D25 may be configured similarly to the media data 407. The media data D25 may be handled similarly to the media data 407, except that the first large-scale generative model M1 is replaced with the second large-scale generative model M2. In one example, the media data D25 may be collected in advance. At least a portion of the collected media data D25 may be provided to the second large-scale generative model M2 as a prompt P21. Alternatively, the prompt P21 may further include a reference instruction I25 that references the media data D25. In response to this, the media data D25 may be referenced by the second large-scale generative model M2 during the calculation process. The reference instruction I25 may include an instruction to search for media data D25. As a result, the media data D25 may be acquired by having the second large-scale generative model M2 search for it. The media data D25 provided to the second large-scale generative model M2 may be narrowed down from all data that can be used as media data D25 to data that indicates social trends at the same or similar time as the time when the satisfaction of notification condition 4 is to be determined. According to one example of this embodiment, by having the second large-scale generative model M2 refer to the media data D25, it is possible to expect an appropriate determination of the satisfaction of notification condition 4 in accordance with social trends.

[0097] Also, in one example, the result 30 of the first analysis may be obtained by a route different from the second large-scale generative model M2. For example, the processing of the first analysis may be performed by a large-scale generative model different from the second large-scale generative model M2. The processing of the first analysis may be performed by a method other than using a large-scale generative model. In this case, the performance of the first analysis by the second large-scale generative model M2 may be omitted. The analysis instruction I21 may be omitted from the prompt P21. The result 30 of the first analysis obtained by a different route may be provided to the second large-scale generative model M2 as the prompt P21 or during the calculation process.

[0098] [Notification condition determination method] The satisfaction of notification condition 4 may be determined directly or indirectly from first analysis result 30. For example, directly determining the satisfaction of notification condition 4 from first analysis result 30 may be configured by determining the satisfaction of notification condition 4 according to at least one of statistics 301 and ranking 303 derived as first analysis result 30. Furthermore, indirectly determining the satisfaction of notification condition 4 from first analysis result 30 may be configured by applying any arithmetic processing (processing, conversion, etc.) to first analysis result 30, and determining the satisfaction of notification condition 4 based on the result of the application of the arithmetic processing.

[0099] 5 schematically illustrates an example of a method for indirectly determining whether notification condition 4 is satisfied according to the present embodiment. In one example, monitoring whether first analysis result 30 satisfies notification condition 4 may include determining whether first analysis result 30 satisfies notification condition 4 by analyzing explanatory data 305 generated from first analysis result 30, the explanatory data 305 explaining first analysis result 30. That is, first analysis result 30 may be converted into explanatory data 305, and whether notification condition 4 is satisfied may be determined according to explanatory data 305. Explanation data 305 may be in any data format, such as text, sound (including audio), or image (diagram, graph, etc.).

[0100] The explanatory data 305 may be generated in any manner from the first analysis results 30. In one example, at least a portion of the explanatory data 305 may be rule-based generated from the first analysis results 30. In another example, at least a portion of the explanatory data 305 may be generated by a trained machine learning model, such as a large-scale generative model.

[0101] For example, as shown in FIG. 5, a third large-scale generative model M3 may be used to generate the explanation data 305. The third large-scale generative model M3 is a large-scale generative model used to generate the explanation data 305. Except for the difference in the purpose of use, the third large-scale generative model M3 may be configured similarly to the first large-scale generative model M1, etc. That is, the type of the third large-scale generative model M3 is not particularly limited and may be appropriately selected depending on the embodiment. The third large-scale generative model M3 may be, for example, a large-scale language model, a large-scale visual language model, a large-scale speech model, etc. The third large-scale generative model M3 may include a multimodal model. The configuration and the number of parameters of the third large-scale generative model M3 may be appropriately determined depending on the embodiment. The third large-scale generative model M3 may be fine-tuned depending on the purpose of use. As a specific example, a known model such as Claude, GPT, or CyberAgentLM may be used for the third large-scale generative model M3. In addition, in one example, the third large-scale generative model M3 may be a large-scale generative model M3. The model M3 may be the same as the first large-scale generative model M1 or the second large-scale generative model M2. In another example, the third large-scale generative model M3 may be a large-scale generative model different from the first large-scale generative model M1 and the second large-scale generative model M2.

[0102] The computational processing of the third large-scale generative model M3 may be executed on any computer. The third large-scale generative model M3 may be deployed on the information processing device 1, and the computational processing of the third large-scale generative model M3 may be executed on the information processing device 1. In another example, the third large-scale generative model M3 may be deployed on a computer other than the information processing device 1, and the computational processing of the third large-scale generative model M3 may be executed on the other computer. When the process of determining whether the notification condition 4 is satisfied based on the explanation data 305 is executed on the information processing device 1, the information processing device 1 may appropriately acquire the explanation data 305 generated by the third large-scale generative model M3 from another computer.

[0103] The instruction given to the third large-scale generative model M3 when generating the explanatory data 305 may be configured appropriately depending on the embodiment. In one example, a prompt P3 including a generation instruction I3 may be given to the third large-scale generative model M3 to cause the third large-scale generative model M3 to generate the explanatory data 305. The prompt P3 may further include at least one of at least a portion of the target data 2 and the result 30 of the first analysis. At least one of at least a portion of the target data 2 and the result 30 of the first analysis may be referenced during the calculation process rather than being given as the prompt P3. As long as it is possible to instruct the generation of the explanatory data 305, the configuration of the generation instruction I3 is not particularly limited and may be determined appropriately depending on the embodiment. In one example, the generation instruction I3 may be given in natural language, such as "Please create a report (explanatory data 305) that explains the analysis results." The data format of the generation instruction I3 is not particularly limited and may be selected appropriately depending on the embodiment. In one example, the format of the generation instruction I3 shown in FIG. 3 or FIG. 4 may be used. In this embodiment, the second large-scale generative model M2 may be caused to execute the process of generating explanation data 305 together with the first analysis process. In this case, the second large-scale generative model M2 corresponds to the third large-scale generative model M3. The generation instruction I3 may be included in the analysis instructions (I20, I21), so that the prompt P3 may be included in the prompts (P20, P21).

[0104] Furthermore, the method of determining whether notification condition 4 is satisfied based on the explanatory data 305 is not particularly limited and may be selected appropriately depending on the embodiment. In one example, whether notification condition 4 is satisfied may be determined from the explanatory data 305 on a rule-based basis. In this case, analyzing the explanatory data 305 may be performed by a rule-based calculation. In another example, whether notification condition 4 is satisfied may be determined from the explanatory data 305 by using a trained machine learning model such as a large-scale generative model. In this case, analyzing the explanatory data 305 may be performed by a calculation of the trained machine learning model. For example, in the embodiment of FIG. 4 above, after the process of generating the explanatory data 305 is performed, the second large-scale generative model M2 may be caused to perform a process of determining whether notification condition 4 is satisfied based on the explanatory data 305. In this case, the second large-scale generative model M2 corresponds to the third large-scale generative model M3.

[0105] In one example of this embodiment, the satisfaction of notification condition 4 is not determined directly from first analysis result 30, but indirectly via explanation data 305. In explanation data 305, the context of first analysis result 30 can be reflected in explanations using text, sound, images, etc. Therefore, according to this example of this embodiment, it can be expected that the explanation data 305 will enable flexible determination according to the context.

[0106] [Notification data] As long as the second analysis result 35 can be transmitted, the configuration of the notification data 5 is not particularly limited and may be determined appropriately depending on the embodiment. In one example, the notification data 5 may be directly configured from the second analysis result 35. For example, the second analysis result 35 may be used as the notification data 5 as is. In another example, any arithmetic processing (processing, conversion, etc.) may be applied to the second analysis result 35. The notification data 5 may be configured from the result of the arithmetic processing together with the second analysis result 35 or instead of the second analysis result 35. The arithmetic processing applied to the second analysis result 35 may include, for example, generating explanatory data (explanation data 355 described below), formatting according to a given template, etc. The notification data 5 may also be referred to as report data, etc.

[0107] 6 schematically illustrates an example of a method for generating notification data 5 according to this embodiment. In one example, the notification data 5 may include search information 55 related to search results obtained by providing a search engine SE with words WD included in at least one of the second analysis results 35, the notification conditions 4, and the explanation data 355. The explanation data 355 may be generated from the second analysis results 35 and may be configured to explain the second analysis results 35.

[0108] The explanatory data 355 may be configured in any data format, such as text, sound (including voice), or images (diagrams, graphs, etc.). The explanatory data 355 may be generated in a manner similar to the explanatory data 305 described above, except that the subject of the explanation is replaced by the second analysis instead of the first analysis. At least a portion of the explanatory data 355 may be generated rule-based from the results 35 of the second analysis, or may be generated by a trained machine learning model. The explanatory data 355 may be configured in a manner similar to the explanatory data 305 described above.

[0109] For example, as shown in FIG. 6, a fourth large-scale generative model M4 may be used to generate the explanatory data 355. The fourth large-scale generative model M4 is a large-scale generative model used to generate the explanatory data 355. Except for the difference in the purpose of use, the fourth large-scale generative model M4 may be configured in the same manner as the first large-scale generative model M1, etc. The fourth large-scale generative model M4 The type is not particularly limited and may be selected appropriately depending on the embodiment. The fourth large-scale generative model M4 may be, for example, a large-scale language model, a large-scale visual language model, a large-scale speech model, etc. The fourth large-scale generative model M4 may include a multimodal model. The configuration and number of parameters of the fourth large-scale generative model M4 may be determined appropriately depending on the embodiment. The fourth large-scale generative model M4 may be fine-tuned depending on the purpose of use. As a specific example, a known model such as Claude, GPT, CyberAgentLM, etc. may be used for the fourth large-scale generative model M4. In one example, the fourth large-scale generative model M4 is The fourth large-scale generative model M4 may be identical to the first large-scale generative model M1, the second large-scale generative model M2, or the third large-scale generative model M3. In another example, the fourth large-scale generative model M4 may be a large-scale generative model different from the first large-scale generative model M1, the second large-scale generative model M2, and the third large-scale generative model M3. The computational processing of the fourth large-scale generative model M4 may be executed on any computer. The fourth large-scale generative model M4 may be deployed on the information processing device 1 or another computer. The computational processing of the fourth large-scale generative model M4 may be executed on the information processing device 1 or another computer.

[0110] The instruction given to the fourth large-scale generative model M4 when generating the explanatory data 355 may be configured appropriately depending on the embodiment. In one example, except for replacing the first analysis with the second analysis, the prompt P4 given to the fourth large-scale generative model M4 may be configured similarly to the prompt P3 given to the third large-scale generative model M3 when generating the explanatory data 305. That is, in one example, a prompt P4 including a generation instruction I4 may be given to the fourth large-scale generative model M4 to cause the fourth large-scale generative model M4 to generate the explanatory data 355. The prompt P4 may further include at least one of at least a portion of the target data 2 and the result 35 of the second analysis. At least one of at least a portion of the target data 2 and the result 35 of the second analysis may be referenced during the calculation process rather than being given as the prompt P4. As long as the generation instruction I4 can instruct an explanation of the explanatory data 355, the configuration of the generation instruction I4 is not particularly limited and may be determined appropriately depending on the embodiment. The generation instruction I4 may be configured similarly to the generation instruction I3 of the prompt P3. In one example, the generation instruction I4 may be given in natural language, such as "Please create a report (explanation data 355) explaining the analysis results." The data format of the generation instruction I4 is not particularly limited and may be selected appropriately depending on the embodiment. When the second analysis is the same as the first analysis, the prompt P3, the third large-scale generative model M3, and the explanation data 305 correspond to the prompt P4, the fourth large-scale generative model M4, and the explanation data 355.

[0111] The word WD to be used as a search query may be selected from words included in the second analysis result 35, the notification condition 4, and the description data 355 using any method (e.g., natural language processing, etc.). The selection may be extraction. The word WD may be appropriately extracted from at least one of the second analysis result 35, the notification condition 4, and the description data 355 using a predetermined method. A specific word used in at least one of the second analysis result 35, the notification condition 4, and the description data 355 may be selected as the word WD. In one example, the word WD may be extracted based on rules such as frequently occurring words or words used in a specific context. The word WD may be extracted randomly. In another example, a trained machine learning model such as a large-scale generative model (e.g., the fourth large-scale generative model M4) may be used to extract the word WD. A known method may be used to extract the word WD. If the word WD is not extracted from the description data 355, the description data 355 may be omitted.

[0112] The search engine SE may be selected arbitrarily. A known engine such as a search engine by Google LLC may be used as the search engine SE. The search scope of the search engine SE need not be particularly limited and may be defined appropriately depending on the embodiment. In one example, the search scope may be defined in advance, such as a specific media (website, etc.), information transmitted during a specific period, etc. The media to be searched may be selected appropriately depending on the embodiment. Another example In the SE, the search range does not need to be specified. The search range may depend on the search engine SE. Search information 55 may be extracted from the top search results as appropriate.

[0113] As long as the search results can be transmitted, the configuration of the search information 55 is not particularly limited and may be determined appropriately depending on the embodiment. In one example, the search information 55 may be directly configured from the search results. For example, the search results may be used as the search information 55 as is. In another example, any arithmetic processing (processing, conversion, etc.) may be applied to the search results. The search information 55 may be configured from the results of the arithmetic processing together with or instead of the search results. The search information 55 may further include any information related to the search other than the search results (e.g., the words WD used in the search, the name of the search engine SE, etc.).

[0114] For example, if the target data 2 includes purchase record data, the type of product may be extracted as a term WD, and information about the type of product may be obtained accordingly as search information 55. The notification data 5 may be configured to include the obtained information. As a specific example, if the product includes beer, the notification data 5 may include information about beer. The information about beer may include, for example, information on social media about target beer products, information on social media about new beer products, news about recent beer sales, etc. As in this example, according to one example of the present embodiment, the inclusion of search information 55 can be expected to provide notification data 5 with rich content.

[0115] (Notification data generation process) The processing procedure for generating the notification data 5 may be determined as appropriate depending on the configuration of the notification data 5. As described above, the configuration of the notification data 5 may be determined as appropriate depending on the embodiment. In one example, the notification data 5 may include at least a portion of the results 35 of the second analysis. The notification data 5 may include explanation data 355 together with or instead of at least a portion of the results 35 of the second analysis. The notification data 5 may also include search information 55 together with or instead of the explanation data 355.

[0116] As an example of a processing procedure, if the notification data 5 includes description data 355, generating the notification data 5 may include performing a second analysis (generating a result of the second analysis 35), generating the description data 355 from the result of the second analysis 35, and formatting the notification data 5 according to a template. If the notification data 5 further includes search information 55, generating the notification data 5 may further include acquiring the search information 55 from a search engine SE. Acquiring the search information 55 may be performed at any timing before formatting. For example, acquiring the search information 55 may be performed after generating the description data 355. If the term WD is not extracted from the description data 355, acquiring the search information 55 may be performed before generating the description data 355. Furthermore, if the term WD is not extracted from the result of the second analysis 35 (if the term WD is extracted from the notification condition 4), acquiring the search information 55 may be performed before performing the second analysis. Formatting may be performed by arranging the components of the notification data 5 to fit a predetermined template (e.g., inserting them into locations specified by the template). The components of the notification data 5 may be the second analysis result 35, the description data 355, the search information 55, etc. The template may be provided in any way. The template may be provided in a CSV (comma The data may be provided in a known manner such as a "separated values" file.

[0117] The processes for generating the notification data 5 (the second analysis process, the generation process of the explanatory data 355, the acquisition process of the search information 55, the formatting process, etc.) may be executed by any computer. In one example, the second analysis process, the generation of the explanatory data 355, the acquisition of the search information 55, and the formatting may be executed on the information processing device 1. The template used for formatting may be stored in the memory resources of the information processing device 1, or may be stored in the information processing device 1. The notification data 5 may be stored in an external storage device accessible to the information processing device 1. In another example, at least one of the second analysis process, the generation of the explanation data 355, the acquisition of the search information 55, and the formatting may be executed on a computer other than the information processing device 1. When the final notification data 5 is generated by the other computer after formatting, and the transmission of the notification data 5 is executed by the information processing device 1, the information processing device 1 may appropriately acquire the notification data 5 generated by the other computer. The generation of the notification data 5 by the information processing device 1 may be configured by having the other computer generate the notification data 5.

[0118] In another example, the process of generating notification data 5 may be executed only if it is determined that the first analysis result 30 satisfies the notification condition 4. In other words, if the first analysis result 30 does not satisfy the notification condition 4, at least part of the process of generating notification data 5 may be omitted. For example, at least one of generating the description data 355, acquiring the search information 55, and formatting may be omitted. If the second analysis differs from the first analysis, the second analysis process may also be omitted. In another example, the process of generating notification data 5 may be executed regardless of whether the first analysis result 30 satisfies the notification condition 4. In other words, even if the first analysis result 30 does not satisfy the notification condition 4 and the notification data 5 is not notified to the user terminal U0, the notification data 5 may be generated. In this case, the notification data 5 that was not notified may be handled arbitrarily. In one example, the generated notification data 5 may be stored as history in an arbitrary storage device, regardless of whether it is notified. The arbitrary storage device may be appropriately selected from, for example, the memory resource MR, the memory resource of the information processing device 1, or another storage device. In this case, the notification data 5 that has not been notified may be provided to the user Z0 as appropriate in response to a request from the user terminal U0 (such as access to a storage device). The notification data 5 may be erased as appropriate. For example, the notification data 5 may be erased upon expiration of the storage period. The notification data 5 that has not been notified may be erased without being stored.

[0119] The process of generating notification data 5 may be interpreted as being included in either the process of monitoring whether notification condition 4 is satisfied or the process of notifying notification data 5. In one example, the process of generating notification data 5 may be interpreted as being included in only one of the process of monitoring whether notification condition 4 is satisfied or the process of notifying notification data 5. In another example, part of the process of generating notification data 5 may be interpreted as being included in the process of monitoring whether notification condition 4 is satisfied and the rest of the process may be interpreted as being included in the process of notifying notification data 5. For example, if the second analysis is the same as the first analysis, the process of the second analysis may be interpreted as being included in the process of monitoring whether notification condition 4 is satisfied or not. If the second analysis is the same as the first analysis and a form is adopted in which whether notification condition 4 is satisfied or not is determined based on description data 305 (description data 355), the process of generating description data 305 (description data 355) may be interpreted as being included in the process of monitoring whether notification condition 4 is satisfied or not. When adopting a form in which notification data 5 is generated only when notification condition 4 is satisfied, the process of acquiring search information 55 and the process of formatting it may be construed as being included in the process of notifying notification data 5. The process of generating notification data 5 may be construed as a process separate from the process of monitoring whether notification condition 4 is satisfied and the process of notifying notification data 5.

[0120] (Notification data notification method) The process of transmitting the notification data 5 to the user terminal U0 may be executed by any computer. In one example, the process of transmitting the notification data 5 may be executed by the information processing device 1. That is, notifying the notification data 5 may include transmitting the notification data 5. For example, the information processing device 1 may generate the notification data 5 and transmit the generated notification data 5 to the user terminal U0. Also, for example, the notification data 5 may be generated by a computer other than the information processing device 1. The information processing device 1 may directly or indirectly acquire the notification data 5 from the other computer and transmit the acquired notification data 5 to the user terminal U0.

[0121] FIG. 7 is a diagram showing another example of a method for notifying the notification data 5 according to the present embodiment. In one example, the process of transmitting the notification data 5 may be executed by a computer EC other than the information processing device 1. The information processing device 1 may cause the other computer EC to transmit the notification data 5 by providing a notification instruction NI9 to the other computer EC. That is, notifying the other computer EC of the notification data 5 may include instructing the other computer EC to transmit the notification data 5 (permitting the transmission of the notification data 5). As long as the notification instruction NI9 can instruct the transmission of the notification data 5, the configuration of the notification instruction NI9 is ​​not particularly limited and may be determined appropriately depending on the embodiment. The notification instruction NI9 may be configured by a transmission permission. For example, the other computer EC may generate the notification data 5. The other computer EC may generate the notification data 5 autonomously or may generate the notification data 5 in response to an instruction from the information processing device 1 or the like. The other computer EC may transmit the generated notification data 5 in response to the notification instruction NI9. Furthermore, for example, the notification data 5 may be generated by the information processing device 1 or another computer. The other computer EC may acquire the notification data 5 from the information processing device 1 or another computer. The other computer EC may transmit the acquired notification data 5 in response to the notification instruction NI9. The user terminal U0 to which the notification data 5 is to be sent may be specified as appropriate depending on the embodiment. The notification destination (user terminal U0) may be specified by, for example, an email address, an account, a telephone number, a terminal ID, etc.

[0122] [Response to notification data] FIG. 8 schematically shows an example of a scene in which user Z0 returns a response 59 to notification data 5 according to this embodiment. In one example, after notifying the information processing device 1 of the notification data 5, the information processing device 1 may accept the response 59 from user Z0 via the user terminal U0. The response 59 is feedback from user Z0. The response 59 may include any request. The information processing device 1 may perform information processing corresponding to the request from user Z0 included in the response 59, and may return the results of the information processing to the user terminal U0.

[0123] (Example request) In one example, the response 59 may include an additional request for the second analysis. By receiving this response 59, the information processing device 1 may perform the second analysis in accordance with the request included in the response 59. The information processing device 1 may generate new notification data 5 from the obtained second analysis result 35. The new notification data 5 may be generated in the same manner as the original notification data 5. The information processing device 1 may notify the user terminal U0 of the generated new notification data 5 in response to the response 59. For example, if the target data 2 includes purchase record data and the original notification data 5 includes a total value of sales by product type (total sales of beer, etc.), the response 59 may include a request for analysis results other than the total value (other statistics, rankings, etc.). In response to this, the information processing device 1 may perform the second analysis again to obtain analysis results other than the total value (second analysis result 35) and may return new notification data 5 including the obtained analysis results to the user terminal U0. For example, if the response 59 includes a request for the sales ranking of each product (such as the sales ranking of each brand of beer), the information processing device 1 may return new notification data 5 including the sales ranking of each product.

[0124] The content of the second analysis requested by the user Z0 may be appropriately identified from the response 59. The identified content may be analysis items such as statistics, rankings, etc. In one example, the information processing device 1 may identify the content of the second analysis from the response 59 on a rule-based basis. In another example, the information processing device 1 may identify the content of the second analysis from the response 59 using a trained machine learning model such as a large-scale generative model. For example, the information processing device 1 may cause the large-scale generative model to determine the content of the second analysis by providing a prompt including the notification data 5, the response 59, and an instruction to the large-scale generative model. In one example, the information processing device 1 may cause the large-scale generative model to generate a processing procedure for the second analysis such as program code. As long as it is possible to instruct the large-scale generative model to determine the content of the second analysis (which may include generating the processing procedure for the second analysis), the configuration of the instruction to be provided to the large-scale generative model is not particularly limited and may be appropriately determined depending on the embodiment. Large-scale generative model The model may be configured similarly to the first large-scale generative model M1, etc. The large-scale generative model may be the same as or different from each of the large-scale generative models (M1, M2, M3, M4) described above. This makes it possible to expect the provision of new notification data 5 that matches the intentions of user Z0.

[0125] In one example, the response 59 may include a request for at least one of the first analysis and the notification condition 4. By receiving this response 59, the information processing device 1 may change at least one of the content of the first analysis and the setting of the notification condition 4 in accordance with the request included in the response 59. As described above, in one example, the information processing device 1 may rule-based identify at least one of the content of the first analysis and the notification condition 4 from the response 59. In another example, the information processing device 1 may use a trained machine learning model, such as a large-scale generative model, to identify at least one of the content of the first analysis and the notification condition 4 from the response 59. The content of the first analysis may be identified in a manner similar to that of the second analysis. Furthermore, the information processing device 1 may generate the notification condition 4 from the response 59 by treating the response 59 in the same manner as the condition specification data 401. This makes it possible to expect at least one of the first analysis and the notification condition 4 to be set in accordance with the intention of the user Z0.

[0126] For example, if the target data 2 includes purchase record data and a threshold for the total sales value of the target product is set as the notification condition 4, the response 59 may include a request to change (increase or decrease) the threshold. The information processing device 1 may change the threshold of the notification condition 4 in response to this response 59. Furthermore, for example, the response 59 may include a request to change the analysis item to be monitored from the total sales value to a ranking. In response to this response 59, the information processing device 1 may change the content of the first analysis to calculate the ranking of sales, and change the setting of the notification condition 4 to determine whether the ranking is sufficient in accordance with the change in ranking.

[0127] (response) Response 59 may be entered in any manner. In one example, response 59 may be entered free-form, such as by text, voice, image, etc.

[0128] In another example, the information processing device 1 may predict one or more reply candidates for the user Z0 from the notification data 5. The one or more reply candidates may be predicted by any method. In one example, the information processing device 1 may predict one or more reply candidates from the notification data 5 based on rules (for example, by selecting frequently requested requests from past history as reply candidates). In another example, the information processing device 1 may predict one or more reply candidates using a trained machine learning model such as a large-scale generative model. The explanatory variables for predicting the one or more reply candidates may include other information (such as notification conditions 4) together with or instead of the notification data 5. The information processing device 1 may notify one or more reply candidates along with the notification data 5. The user terminal U0 may output one or more reply candidates along with the notification data 5. The user Z0 may input a reply 59 by selecting a desired reply candidate from one or more reply candidates. In this case, the selected reply candidate is an example of the reply 59. This is expected to reduce the effort required by the user Z0 to input the reply 59. For example, if the target data 2 includes purchase record data and the original notification data 5 includes the total sales value of a product type (e.g., total sales of beer), a request for analysis results other than the total value (e.g., other statistics, rankings, etc.) may be predicted as a candidate response. In response to this, candidate responses such as a request to calculate other feature quantities or a request to calculate rankings may be output along with the notification data 5. When the user Z0 selects a candidate response indicating a ranking calculation request (receives a response 59 including a request for the sales rankings of each product), the information processing device 1 may acquire the sales rankings of each product, as in the above example. Then, the information processing device 1 may return new notification data 5 including the acquired sales rankings of each product.

[0129] Note that if more than one candidate response is predicted, the candidate responses will be selected as responses 59. In anticipation, the information processing device 1 may execute information processing corresponding to at least one of the predicted one or more candidate answers before accepting the answer 59. In one example, the information processing device 1 may notify the user terminal U0 of the result of the information processing corresponding to the candidate answers together with the notification data 5. The result of the information processing may be output in response to the actual selection of the corresponding candidate answer, or may be output regardless of the selection of the candidate answer. In another example, the information processing device 1 may transmit the result of the information processing corresponding to the candidate answer selected as the answer 59 to the user terminal U0 as a reply to the answer 59 from the user Z0.

[0130] For example, in the case of a request for an addition to the second analysis, the information processing device 1 may perform a second analysis in response to the request for an addition indicated by a candidate answer. The information processing device 1 may notify the user terminal U0 of data indicating a result 35 of the second analysis in response to the request for an addition, together with the notification data 5. The user terminal U0 may output the data indicating the result 35 of the second analysis in response to the request for an addition in response to the selection of a corresponding candidate answer, or may output the data regardless of the selection of a candidate answer. Alternatively, the information processing device 1 may notify the user terminal U0 of new notification data 5 indicating the result 35 of the second analysis corresponding to the candidate answer selected as the response 59, as a reply to the response 59 from the user Z0. Furthermore, for example, in the case of a request for at least one of the first analysis and the notification condition 4, the information processing device 1 may specify changes to at least one of the first analysis and the notification condition 4 in response to the request indicated by the candidate answer. The information processing device 1 may execute changes to the specified content corresponding to the selected candidate answer in response to the selection of a candidate answer by the user Z0.

[0131] In addition, in one example, the information processing device 1 may determine whether to provide a candidate answer to the user Z0 depending on the result of information processing corresponding to the candidate answer. The method of determination may be defined arbitrarily. For example, in the case of a request for additional second analysis described above, the information processing device 1 may cause the large-scale generative model to determine whether to provide a candidate answer to the user Z0 by providing the large-scale generative model with a prompt including the candidate answer (request for the second analysis), the corresponding result 35 of the second analysis, and an instruction. A candidate answer determined not to be provided may be excluded from the notification targets. This is expected to reduce the amount of data notified to the user terminal U0.

[0132] (others) In the above embodiment, the information processing for the response 59 from the user Z0 (acceptance of the response 59, generation of new notification data 5, prediction of response candidates, reply to the response 59, etc.) is described as being executed on the information processing device 1. However, the entity that executes the information processing for the response 59 is not limited to this example and may be changed as appropriate depending on the embodiment. At least a part of the information processing for the response 59 may be executed by another computer (another computer EC, etc.) other than the information processing device 1. The acceptance of the response 59 and the reply to the response 59 may be repeated multiple times. Furthermore, the acceptance of the response 59 may be omitted. In another example, the information processing regarding the provision of the notification data 5 may be completed when the notification of the notification data 5 is completed.

[0133] [Task Completion] 9 schematically illustrates an example of a scene in which task 7 according to this embodiment is performed. In one example, the information processing device 1 may monitor whether a result 39 of a third analysis of target data 2 that is continuously collected satisfies a task execution condition 6. If the result 39 of the third analysis satisfies the task execution condition 6 as a result of the monitoring, the information processing device 1 may perform a predetermined task 7.

[0134] The third analysis may be configured in the same manner as the first and second analyses. The third analysis may include statistical analysis. The calculation process of the third analysis may be defined by any mathematical formula. A trained machine learning model may be used in the third analysis. The range of the target data 2 to be processed by the third analysis may be determined appropriately depending on the embodiment. The third analysis may be performed based on the task execution conditions 6. The third analysis is performed to determine whether a predetermined condition is satisfied. In one example, the third analysis may be identical (common) to the first analysis or the second analysis. In one example, the third analysis may be included in the first analysis or the second analysis as a part of the first analysis or the second analysis. In one example, the third analysis may include the first analysis or the second analysis as a part. In one example, the third analysis may partially coincide with the first analysis or the second analysis. In one example, the third analysis may differ from the first analysis or the second analysis. The third analysis may be performed at any timing. In one example, like the first analysis, the third analysis may be performed in response to the satisfaction of a predetermined condition. If the third analysis is identical to the first analysis, analysis processes corresponding to the first analysis and the third analysis may be performed in response to the satisfaction of a predetermined condition.

[0135] The task execution condition 6 is a trigger condition for determining whether to perform a predetermined task 7. The task execution condition 6 may be configured similarly to the notification condition 4, except that the first analysis is replaced with the third analysis. In one example, the task execution condition 6 may include a change in the trend of the target data 2 in the result 39 of the third analysis. In one example, the third analysis may include calculating statistics of the target data 2. The change in trend may be evaluated according to the calculated statistics. In another example, the target data 2 may include numerical data for each of a plurality of items. The third analysis may include calculating a ranking of each of the plurality of items from the numerical data. The change in trend may be evaluated according to a change in the calculated ranking. Monitoring the fulfillment of the task execution condition 6 may include determining whether the task execution condition 6 is fulfilled. Monitoring (determining) the fulfillment of the task execution condition 6 may be performed continuously. Monitoring the fulfillment of the task execution condition 6 may be performed at any timing.

[0136] As long as the sufficiency can be determined, the data format of the task execution conditions 6 is not particularly limited and may be appropriately selected depending on the embodiment. For example, at least a portion of the task execution conditions 6 may be defined in a numerical format, such as a threshold value. For example, at least a portion of the task execution conditions 6 may be defined in a data format other than numerical values, such as text. Furthermore, the method for setting the task execution conditions 6 is not particularly limited and may be appropriately selected depending on the embodiment. The task execution conditions 6 may be set in a manner similar to the notification conditions 4 described above. For example, at least a portion of the task execution conditions 6 may be provided in advance. At least a portion of the task execution conditions 6 may be predefined within a program. For example, at least a portion of the task execution conditions 6 may be manually set by an operator. For example, at least a portion of the task execution conditions 6 may be automatically set at least in part by information processing, such as using a large-scale generative model. At least a portion of the task execution conditions 6 may be generated from reference data related to the task execution conditions 6 by a large-scale generative model (such as the first large-scale generative model M1). The reference data may be configured similarly to the reference data 40 used to generate the notification conditions 4 described above.

[0137] The predetermined task 7 is not particularly limited and may be defined appropriately depending on the embodiment. The predetermined task 7 may include any type of task that can be executed by information processing. The predetermined task 7 may include control of a device such as a robotic device. In one example, the predetermined task 7 may include notifying notification data 5. In this case, the first analysis and notification condition 4 correspond to the third analysis and task execution condition 6. In another example, the predetermined task 7 may include any information processing other than notifying notification data 5. The content of the third analysis and the task execution condition 6 may be set to be related to the predetermined task 7.

[0138] For example, if the target data 2 includes sensing data, the third analysis may include analyzing the situation (trend, etc.) of the object being observed by the sensor. The task execution condition 6 may be set to relate to the situation of the object being observed, such as, for example, vital signs satisfying a predetermined condition, or the situation of the house satisfying a predetermined condition. The predetermined task 7 may be any task according to the situation of the object being observed that appears in the sensing data. As a specific example, the sensor may be a blood glucose meter, and the sensing data may be measurement data of blood glucose levels. The task execution condition 6 may be set to relate to the situation of the object being observed, such as, for example, the vital signs satisfying a predetermined condition, or the situation of the house satisfying a predetermined condition. The task execution condition 6 may include whether or not there is anyone at home. Whether or not there is anyone at home may be determined by any method, such as threshold comparison. The predetermined task 7 may be to output a notification instructing the administration of insulin. The predetermined task 7 may be to instruct an insulin pump to administer insulin. As a specific example, the sensor may be an IoT sensor used in the home, and the sensing data may be measurement data from the IoT sensor. The task execution condition 6 may include whether or not there is anyone at home. Whether or not there is anyone at home may be determined by any method. The predetermined task 7 may be to turn off the power of the target device when there is no one at home. The target device may be any device whose power supply can be controlled.

[0139] For example, if the target data 2 includes financial data, the third analysis may include analyzing the financial situation (trends, etc.). The task execution condition 6 may be set to relate to the financial situation. The predetermined task 7 may be any task corresponding to the financial situation expressed in the financial data. As a specific example, the financial data may be stock price data, and the third analysis (analysis of the financial situation) may include calculating stock price fluctuations. The task execution condition 6 may include the calculated stock price fluctuations satisfying predetermined conditions (such as stop-loss criteria). The predetermined task 7 may include buying and selling stocks in response to the fulfillment of predetermined conditions, such as selling stocks when stop-loss criteria are met. The buying and selling of stocks may be performed using known information processing.

[0140] Furthermore, for example, if the target data 2 includes payment data, the third analysis may include analyzing the payment status (trends, etc.). The task execution condition 6 may be set to relate to the payment status. The predetermined task 7 may be any task depending on the payment status expressed in the payment data. As a specific example, the payment data may be credit card payment data, and the third analysis (analysis of the payment status) may include detecting whether at least one of the amount and the purchased items deviates from the purchase history. The task execution condition 6 may include at least one of the amount and the purchased items deviating from the purchase history. The predetermined task 7 may include notifying the user terminal U0 of a warning that the credit card may be being used by another person if at least one of the amount and the purchased items deviates from the purchase history.

[0141] The third analysis, the determination of whether the task execution condition 6 is satisfied, and the processing of the predetermined task 7 may be performed on any computer. In one example, the third analysis, the determination of whether the task execution condition 6 is satisfied, and the processing of the predetermined task 7 may be performed on the information processing device 1. In another example, at least one of the third analysis, the determination of whether the task execution condition 6 is satisfied, and the processing of the predetermined task 7 may be performed on a computer other than the information processing device 1. If the determination of whether the task execution condition 6 is satisfied is performed on another computer, monitoring the satisfaction of the task execution condition 6 may be configured by obtaining a determination result of whether the task execution condition 6 is satisfied from the other computer. If the processing of the predetermined task 7 is performed on another computer, performing the predetermined task 7 may be configured by instructing the other computer to perform the predetermined task 7.

[0142] The process of determining whether the task execution condition 6 is satisfied may be defined by any mathematical expression. In one example, the process of determining whether the task execution condition 6 is satisfied may be configured based on a rule. In one example, the process of determining whether the task execution condition 6 is satisfied may be configured by program code generated by a large-scale generative model. A trained machine learning model such as a large-scale generative model may be used for the process of determining whether the task execution condition 6 is satisfied. In one example, monitoring whether the result 39 of the third analysis satisfies the task execution condition 6 may include causing the large-scale generative model to perform the third analysis by providing at least a portion of the target data 2 as input. The embodiment of FIG. 3 may also be employed for performing the third analysis. In one example, monitoring whether the result 39 of the third analysis satisfies the task execution condition 6 may include causing the large-scale generative model to perform the third analysis by providing at least a portion of the target data 2 as input, and 4 may also be used to determine whether the task execution condition 6 is satisfied by replacing the matters related to the notification condition 4 with the task execution condition 6.

[0143] The satisfaction of the task execution condition 6 may be determined directly or indirectly from the third analysis result 39. In one example, monitoring whether the third analysis result 39 satisfies the task execution condition 6 may include determining whether the third analysis result 39 satisfies the task execution condition 6 by analyzing explanatory data generated from the third analysis result 39, the explanatory data explaining the third analysis result 39. The method for determining whether the task execution condition 6 is satisfied using the explanatory data may be configured similarly to that shown in FIG. 5 above. The explanatory data may be generated in a similar manner to the explanatory data 305, etc.

[0144] According to one example of this embodiment, a predetermined task 7 can be performed in a data-driven manner. Therefore, when performing the predetermined task 7, it is expected that the effort required for the user (user Z0) to monitor the data to be analyzed (target data 2) can be reduced. For example, in the case of the sensing data, a task can be performed in a data-driven manner according to the status of the object observed by the sensor (such as issuing an instruction to administer insulin or turning off the power of the target device). In the case of the financial data, a task can be performed in a data-driven manner according to the financial status (such as buying and selling stocks). In the case of the payment data, a task can be performed in a data-driven manner according to the payment status (such as issuing a warning).

[0145] §2 Configuration example [Hardware configuration] 10 is a diagram illustrating an example of a hardware configuration of the information processing device 1 according to this embodiment. In one example, the information processing device 1 may be configured as a computer in which a control unit 11, a storage unit 12, a communication module 13, an input device 14, and an output device 15 are electrically connected.

[0146] The control unit 11 is configured to execute information processing based on programs and various data. The control unit 11 may include a CPU, which is a hardware processor, a RAM (Random Access Memory), a ROM (Read Only Memory), etc. The control unit 11 (CPU) is an example of a processor resource.

[0147] The storage unit 12 is configured to hold any data. The storage unit 12 may include, for example, a hard disk drive, a solid state drive, a semiconductor memory, etc. The storage unit 12, RAM, and ROM are examples of memory resources of the information processing device 1. In one example of this embodiment, the storage unit 12 may store various information such as a program 81. The program 81 is a program for causing the information processing device 1 to execute information processing relating to notification control of the notification data 5 (see Figures 12, 13, and 14 described below). The program 81 includes a series of instructions for the information processing.

[0148] In one example, the program 81 may be stored in a storage medium 91 instead of or together with the storage unit 12. The storage medium 91 is configured to store various types of information (such as the stored program) by electrical, magnetic, optical, mechanical, or chemical action so that a machine such as a computer can read the information. The storage unit 12 and the storage medium 91 are examples of non-transitory storage media. The information processing device 1 may acquire the program 81 from the storage medium 91. The storage medium 91 may be a disk-type storage medium (CD, DVD, etc.) or a non-disk type storage medium such as a semiconductor memory (flash memory, etc.). Any drive device may be used to read information stored in the storage medium 91. The type of drive device may be selected depending on the storage medium 91. The drive device may be connected to the information processing device 1 in any manner. The storage medium 91 may include an external storage device. In one example, when the storage unit 12 is used as a memory resource MR, the collected target data 2 may be stored in the storage unit 12.

[0149] The communication module 13 is configured to perform wired or wireless communication via a network. The communication module 13 may be configured, for example, by a wired LAN (Local Area Network) module, a wireless LAN module, or the like. The network standard is not particularly limited and may be selected appropriately depending on the embodiment. For example, the type of network may be selected appropriately from the Internet, a wireless communication network, a mobile communication network, a telephone network, a dedicated network, or the like. The information processing device 1 may perform data communication with other computers (user terminal U0, other computer EC, target device, etc.) via the communication module 13.

[0150] The input device 14 is configured to accept input of information. The input device 14 may be configured, for example, by a mouse, a keyboard, an operator, etc. The output device 15 is configured to output information. The output device 15 may be configured, for example, by a display, a speaker, etc. An operator can operate the information processing device 1 by using the input device 14 and the output device 15. The input device 14 and the output device 15 may be directly connected to the information processing device 1, or may be indirectly connected via at least one of the communication module 13 and an external interface. The external interface may be appropriately configured to connect to an external device via a wired or wireless connection, for example, by a USB (Universal Serial Bus) port, a dedicated port, etc. The input device 14 and the output device 15 may be at least partially integrated into a touch panel display, etc.

[0151] It should be noted that, with regard to the specific hardware configuration of the information processing device 1, components may be omitted, replaced, or added as appropriate depending on the embodiment. For example, the control unit 11 may include multiple hardware processors. The hardware processors may be a microprocessor, a field-programmable gate array (FPGA), a digital signal processor (DSP), a GP The input device 14 and the output device 15 may be implemented by a graphics processing unit (U), an application specific integrated circuit (ASIC), etc. The program 81 may be omitted. The program 81 may be stored in an external storage device such as a NAS (Network Attached Storage). An external storage device is also an example of a non-transitory storage medium. The information processing device 1 may be configured with multiple computers. In this case, the hardware configuration of each computer may or may not be the same. The information processing device 1 may be a computer designed specifically for the service provided, as well as a general-purpose server device, a general-purpose PC (Personal Computer), a notebook PC, a terminal device, etc. The terminal device may be a smart The user terminal may include a phone, a tablet terminal, or the like.

[0152] [Software configuration] 11 schematically shows an example of the software configuration of the information processing device 1 according to this embodiment. The control unit 11 of the information processing device 1 executes instructions included in the program 81 stored in the storage unit 12 using the CPU. As a result, the information processing device 1 operates as a computer including a monitoring unit 111 and a notification processing unit 112 as software modules. That is, in this example of this embodiment, each software module of the information processing device 1 is realized by the control unit 11 (CPU).

[0153] The monitoring unit 111 is configured to monitor whether or not a result 30 of a first analysis of the target data 2, which is continuously collected, satisfies a notification condition 4. The notification processing unit 112 is configured to notify the user terminal U0 of notification data 5 indicating a result 35 of a second analysis of the target data 2 in response to the result of monitoring indicating that the result 30 of the first analysis satisfies the notification condition 4.

[0154] In one example, the setting process of the notification condition 4 (or the task execution condition 6) is executed on the information processing device 1. In this case, the information processing device 1 may further include a condition setting unit 113 as a software module. The condition setting unit 113 may be configured to determine the notification condition 4 by any method and set the determined notification condition 4. In the case where the form of FIG. 9 (task execution) is adopted, the condition setting unit 113 may be configured to determine the task execution condition 6 by any method and set the determined task execution condition 6.

[0155] In one example, when the processing of the first analysis, the second analysis, or the third analysis is executed on the information processing device 1, the information processing device 1 may further include an analysis processing unit 114 as a software module. The analysis processing unit 114 may be configured to perform the first analysis to generate a result 30 of the first analysis. The analysis processing unit 114 may be configured to perform the second analysis to generate a result 35 of the second analysis. The analysis processing unit 114 may be configured to perform the third analysis to generate a result 39 of the third analysis.

[0156] In one example, when the form of FIG. 9 (task execution) is adopted, the information processing device 1 may further include a task execution unit 115 as a software module. The monitoring unit 111 may be configured to monitor whether or not a result 39 of a third analysis of the continuously collected target data 2 satisfies a task execution condition 6. The task execution unit 115 may be configured to execute a predetermined task 7 in response to the result of monitoring that the result 39 of the third analysis satisfies the task execution condition 6. When the predetermined task 7 is a notification of notification data 5, the notification processing unit 112 corresponds to the task execution unit 115.

[0157] In the example of the present embodiment, each software module of the information processing device 1 is implemented by a general-purpose CPU. However, the method for implementing each module is not limited to this example. Some or all of the software modules may be implemented by one or more dedicated processors or chipsets. Each module may be implemented as a hardware module. With regard to the software configuration of the information processing device 1, modules may be omitted, replaced, or added as appropriate depending on the embodiment.

[0158] §3 Example of operation 12, 13, and 14 show an example of a processing procedure of the information processing device 1 according to this embodiment. Each of the following processing procedures is an example of an information processing method according to the present disclosure. However, each of the following processing procedures is merely an example, and each step may be changed as much as possible. Furthermore, steps in the following processing procedures may be omitted, replaced, or added as appropriate depending on the embodiment.

[0159] [Notification condition settings] Fig. 12 is a flowchart showing an example of a processing procedure for setting a notification condition 4 by the information processing device 1 according to this embodiment. The information processing of Fig. 12 may be executed at any timing. In one example, the information processing device 1 may execute the information processing of Fig. 12 in response to a request from a user terminal (such as the user terminal U1). In another example, the information processing device 1 may autonomously execute the information processing of Fig. 12.

[0160] (Step S901) In step S901, the control unit 11 operates as the condition setting unit 113. As a result, the control unit 11 accepts designation of the notification condition 4 from a user (such as user Z1) via the user terminal. In one example, the control unit 11 may directly accept the designation of the notification condition 4, such as selection of a parameter (such as the statistical quantity 301 or the rank 303) or designation of a parameter value (such as a threshold value). In another example, the control unit 11 may indirectly accept the designation of the notification condition 4, such as selection of the data to be used as the condition designation data 401 or the reference data 40. Upon accepting the designation of the notification condition 4, the control unit 11 proceeds to the next step S902.

[0161] (Steps S902 and S903) In steps S902 and S903, the control unit 11 also operates as the condition setting unit 113. As a result, in step S902, the control unit 11 determines notification condition 4 in accordance with the user's designation in step S901. In step S903, the control unit 11 sets the determined notification condition 4.

[0162] In one example, when the notification condition 4 is directly specified, determining the notification condition 4 may be set to the specified content. Setting the notification condition 4 may include, for example, updating the parameters of the notification condition 4 to selected parameters, updating the values ​​of the parameters to specified values, etc.

[0163] In another example, when the notification conditions 4 are indirectly specified, the control unit 11 may appropriately determine the notification conditions 4 from the specified content. For example, the control unit 11 may cause the first large-scale generative model M1 to determine (generate) at least a portion of the notification conditions 4 by providing reference data 40 to the first large-scale generative model M1. The reference data 40 may include at least one of condition specification data 401, metadata 403, at least a portion of the target data 2, a sample 405, and media data 407. When the condition specification data 401 is obtained in step S901, the control unit 11 may provide the condition specification data 401 to the first large-scale generative model M1 as the reference data 40. The control unit 11 may appropriately reflect the result of determination (generation) by the first large-scale generative model M1 in the setting of the notification conditions 4.

[0164] In this embodiment, the condition specification data 401 is not acquired in real time, and the control unit 11 configures the reference data 40 using data that can be acquired autonomously (pre-acquired condition specification data 401, metadata 403, etc.), so that the notification condition 4 can be determined without any specification by the user. Therefore, in another example, the process of step S901 (the user's specification of the notification condition 4) may be omitted, and the control unit 11 may execute the processes of steps S902 and S903 based on autonomously acquired data.

[0165] When the setting of notification condition 4 is completed, the control unit 11 ends the processing procedure for setting notification condition 4 according to this operation example. Note that, in one example, when the form of FIG. 9 (task execution) is adopted, the control unit 11 operates as the condition setting unit 113, and may set task execution condition 6 in the same procedure as that for notification condition 4.

[0166] [Notification data notification] 13 is a flowchart showing an example of a processing procedure for notifying the analysis result (notification data 5) by the information processing device 1 according to this embodiment. The information processing of FIG. 13 may be executed at any timing. In one example, the first analysis may be executed in response to the satisfaction of a predetermined condition, such as the arrival of a specific time, the passage of a predetermined period, the addition of an amount of data exceeding a threshold, or the addition of unknown data. The information processing of FIG. 13 may be executed in response to the first analysis being performed and the first analysis result 30 being obtained.

[0167] (Steps S101 and S102) In step S101, the control unit 11 operates as the monitoring unit 111. As a result, the control unit 11 monitors whether the first analysis result 30 of the target data 2, which is continuously collected, satisfies the notification condition 4 or not.

[0168] In one example, the first analysis and the determination of whether the notification condition 4 is satisfied may be executed on the information processing device 1. In this case, the control unit 11 may operate as the analysis processing unit 114 and perform the first analysis to generate the result 30 of the first analysis. The control unit 11 may then 0 may determine whether or not the result 30 of the first analysis satisfies the notification condition 4. As a result, the control unit 11 may acquire a determination result of whether the result 30 of the first analysis satisfies the notification condition 4. In another example, at least one of the first analysis and the determination of whether the result 30 of the first analysis satisfies the notification condition 4 may be executed on another computer. When the determination of whether the result 30 of the first analysis satisfies the notification condition 4 is executed on another computer, the control unit 11 may acquire a determination result of whether the result 30 of the first analysis satisfies the notification condition 4 from the other computer.

[0169] In one example, the control unit 11 or another computer may cause the second large-scale generative model M2 to perform a first analysis by providing at least a portion of the target data 2 as input. In one example, the control unit 11 or another computer may cause the second large-scale generative model M2 to perform the first analysis by providing at least a portion of the target data 2 as input, and may also cause the second large-scale generative model M2 to determine whether or not the result 30 of the first analysis satisfies the notification condition 4. In one example, the control unit 11 or another computer may cause the second large-scale generative model M2 to determine whether or not the result 30 of the first analysis satisfies the notification condition 4 by referring to media data D25 related to social trends related to the target data 2. In one example, the control unit 11 or another computer may determine whether or not the result 30 of the first analysis satisfies the notification condition 4 by analyzing the explanation data 305. The first analysis may be performed at any timing before determining whether or not the notification condition 4 is satisfied. In a case where the second large-scale generative model M2 is used, for example, the first analysis may be performed as a series of processes together with the determination of whether or not the notification condition 4 is satisfied.

[0170] In one example, the notification condition 4 may include a change in the trend of the target data 2 in the result 30 of the first analysis. For example, the first analysis may include calculating a statistic 301 of the target data 2. The change in trend may be evaluated according to the calculated statistic 301. Also, for example, the target data 2 may include numerical data 20 for each of a plurality of items. The first analysis may include calculating a rank 303 of each of the plurality of items from the numerical data 20. The change in trend may be evaluated according to the degree of fluctuation in the calculated rank 303. In one example, at least a portion of the notification condition 4 may be generated from the reference data 40 by the first large-scale generative model M1 in the process of step S902 above.

[0171] In step S102, the control unit 11 determines where to branch the process based on the monitoring result of step S101 (the determination result of whether notification condition 4 is satisfied). If the monitoring result shows that first analysis result 30 satisfies notification condition 4, the control unit 11 proceeds to the next step S103. On the other hand, if first analysis result 30 does not satisfy notification condition 4, the control unit 11 skips step S103 and proceeds to step S104.

[0172] (Step S103) In step S103, the control unit 11 operates as the notification processing unit 112. As a result, the control unit 11 notifies the user terminal U0 of the notification data 5 indicating the result 35 of the second analysis on the target data 2.

[0173] The notification data 5 may be appropriately configured to indicate the results of the second analysis 35. In one example, the notification data 5 may include at least one of (at least a portion of) the results of the second analysis 35 and explanatory data 355. In one example, the notification data 5 may include search information 55 regarding search results obtained by providing a search engine SE with words WD included in at least one of the results of the second analysis 35, the notification conditions 4, and the explanatory data 355.

[0174] In one example, the generation of the notification data 5 and the transmission of the notification data 5 may be executed on the information processing device 1. In this case, the control unit 11 may operate as the analysis processing unit 114 and perform the second analysis to generate the result 35 of the second analysis. The control unit 11 may generate the notification data 5 from the generated result 35 of the second analysis. For example, if the notification data 5 is the result 35 of the second analysis, The control unit 11 may generate the explanation data 355 from the second analysis result 35, if the result includes explanation data 355 and search information 55, and if there is a possibility that the term WD used to obtain the search information 55 may be extracted from the explanation data 355. The control unit 11 may extract the term WD from at least one of the second analysis result 35, the notification condition 4, and the explanation data 355. The control unit 11 may obtain the search information 55 by providing the extracted term WD to the search engine SE. The control unit 11 may format the notification data 5 by arranging the second analysis result 35, the explanation data 355, and the search information 5 so that it conforms to a predetermined template. In this way, the control unit 11 may generate the notification data 5. The control unit 11 may transmit the generated notification data 5 to the user terminal U0. In another example, at least one of generating the notification data 5 and transmitting the notification data 5 may be performed on another computer (another computer EC, etc.). When the transmission of the notification data 5 is executed on the other computer EC, the control unit 11 may cause the other computer EC to transmit the notification data 5 by giving a notification instruction NI9 to the other computer EC.

[0175] In one example, the result 35 of the second analysis may be generated by a large-scale generative model (second large-scale generative model M2), similar to the first analysis. Not only the result 35 of the second analysis, but also the notification data 5 itself may be generated by a large-scale generative model. The process of generating the notification data 5 (including the implementation of the second analysis) may be executed at any timing before notifying the notification data 5. After notifying the notification data 5, the control unit 11 proceeds to the next step S104.

[0176] (Step S104) In step S104, the control unit 11 determines whether or not to terminate the processing. The criteria for the determination may be set arbitrarily. In one example, the control unit 11 may determine not to terminate the processing until an instruction to terminate is given. On the other hand, when an instruction to terminate is given, the control unit 11 may determine to terminate the processing. The instruction to terminate may be given in any manner, for example, by terminating / suspending an application.

[0177] If it is determined not to end the processing, the control unit 11 returns the processing to step S101 and executes the processing again from step S101. As a result, the control unit 11 may continuously execute the processing related to the notification of the notification data 5. By repeatedly executing the processing of step S101, the control unit 11 may continuously monitor whether the first analysis result 30 satisfies the notification condition 4. On the other hand, if it is determined to end the processing, the control unit 11 ends the processing procedure related to the notification of the notification data 5 according to this operation example. Note that the timing of ending the processing is not limited to this example. The control unit 11 may end this processing procedure at any timing. Also, in one example, the control unit 11 may execute a series of processes from step S101 to step S104 in real time. Furthermore, in another example, the control unit 11 may receive a response 59 from the user Z0 after notifying the notification data 5 in step S103. The control unit 11 may execute information processing corresponding to the request of the user Z0 included in the response 59 and may return the result of the information processing to the user terminal U0. In another example, at least one of receiving the response 59, executing information processing corresponding to the request, and returning the results of the information processing may be executed on another computer.

[0178] [Task Completion] 14 is a flowchart showing an example of a processing procedure for performing a predetermined task 7 by the information processing device 1 according to this embodiment. The information processing of FIG. 14 may be performed at any timing. In one example, similar to the first analysis, the third analysis may be performed in response to the satisfaction of a predetermined condition, such as the arrival of a specific time, the passage of a predetermined period, the addition of an amount of data exceeding a threshold, or the addition of unknown data. The information processing of FIG. 13 may be performed in response to the third analysis being performed and a result 39 of the third analysis being obtained.

[0179] (Steps S301 and S302) In step S301, the control unit 11 operates as the monitoring unit 111. As a result, the control unit 11 monitors whether or not a result 39 of a third analysis of the target data 2, which is continuously collected, satisfies the task execution condition 6. The processing of step S301 may be configured similarly to step S101, except that the first analysis is replaced with the third analysis and the notification condition 4 is replaced with the task execution condition 6. In one example, the third analysis may be performed on the information processing device 1. In this case, the control unit 11 may operate as the analysis processing unit 114. As a result, the control unit 11 may perform the third analysis to generate a result 39 of the third analysis.

[0180] In step S302, the control unit 11 determines where to branch the process based on the monitoring result of step S301 (the determination result of whether the task execution condition 6 is satisfied). If the monitoring result shows that the third analysis result 39 satisfies the task execution condition 6, the control unit 11 proceeds to the next step S303. On the other hand, if the third analysis result 39 does not satisfy the task execution condition 6, the control unit 11 skips the process of step S303 and proceeds to step S304.

[0181] (Step S303) In step S303, the control unit 11 operates as the task execution unit 115. As a result, the control unit 11 executes the predetermined task 7. In one example, the control unit 11 may execute information processing corresponding to the predetermined task 7. In another example, the control unit 11 may give an instruction to another computer to cause the other computer to execute information processing corresponding to the predetermined task 7. After executing the predetermined task 7, the control unit 11 proceeds to the next step S304.

[0182] (Step S304) In step S304, the control unit 11 determines whether or not to end the process. The process of step S304 may be configured in the same manner as step S104 above.

[0183] If it is determined not to end the processing, the control unit 11 returns the processing to step S301 and executes the processing again from step S301. As a result, the control unit 11 may continuously execute the processing related to the execution of the predetermined task 7. By repeatedly executing the processing of step S301, the control unit 11 may continuously monitor whether the third analysis result 39 satisfies the task execution condition 6. On the other hand, if it is determined to end the processing, the control unit 11 ends the processing procedure related to the execution of the predetermined task 7 according to this operation example. Note that the timing of ending the processing is not limited to this example. The control unit 11 may end this processing procedure at any timing. Also, in one example, the control unit 11 may execute the series of processing from step S301 to step S304 in real time.

[0184] [Features] In this embodiment, the processing in steps S101 and S102 is triggered by the first analysis result 30 of the target data 2 satisfying the notification condition 4. This makes it possible to provide notification data 5 in step S103 in a data-driven manner. Therefore, according to this embodiment, it is expected that the effort required for the user to monitor the target data 2 will be reduced.

[0185] §4 Variations Although the embodiments of the present disclosure have been described in detail above, the above description is merely an example of the present disclosure in every respect. The processes and means described in the present disclosure can be freely combined and implemented as long as no technical contradiction occurs. Furthermore, various improvements or modifications may be made to the above embodiments as appropriate. For example, the following modifications are possible. Note that, in the following, the same reference numerals are used for components similar to those in the above embodiments, and the above Explanations of points similar to those in the above embodiment will be omitted where appropriate.

[0186] <4.1> At least one of omission, replacement, and addition of steps may be performed for each processing procedure of the information processing device 1 according to the above embodiment. For example, the processes of steps S901 to S903 may be omitted. Accordingly, the condition setting unit 113 may be omitted from the software configuration of the information processing device 1. For example, if the first analysis, the second analysis, and the third analysis are not executed on the information processing device 1, the analysis processing unit 114 may be omitted from the software configuration of the information processing device 1. Furthermore, for example, information processing related to the execution of a predetermined task 7 (the processes of steps S301 to S304) may be omitted. Accordingly, the task execution unit 115 may be omitted from the software configuration of the information processing device 1.

[0187] <4.2> In the above embodiment, the configuration related to the execution of the predetermined task 7 may be omitted, but the configuration related to the notification of the notification data 5 may be omitted. In this case, the information processing related to the notification of the notification data 5 (the processing of steps S101 to S104) may be omitted. Accordingly, the notification processing unit 112 may be omitted from the software configuration of the information processing device 1. The first analysis and the second analysis may be omitted. The third analysis may simply be referred to as analysis.

[0188] In this modification, the processing in steps S301 and S302 is triggered by the result 39 of the analysis (third analysis) of the target data 2 satisfying the task execution condition 6. This makes it possible to perform the predetermined task 7 in step S303 in a data-driven manner. Therefore, according to this modification, it is expected that the effort required for the user to monitor the target data 2 when performing the predetermined task 7 can be reduced.

[0189] <4.3> In the above embodiment, the user terminal U0 may be any computer different from the information processing device 1. The user terminal U0 may be appropriately connected to the information processing device 1 via a network. The user terminal U0 may be, for example, a general-purpose PC, a notebook PC, a terminal device, etc. However, the form of the user terminal U0 is not limited to such examples. In another example, the information processing device 1 may be the same computer as the user terminal U0. In this case, notification to the user terminal U0 may be configured by outputting to the output device 15.

[0190] Similarly, in the above embodiment, the user terminal U1 may be any computer different from the information processing device 1. The user terminal U1 may be appropriately connected to the information processing device 1 via a network. The user terminal U1 may be, for example, a general-purpose PC, a notebook PC, a terminal device, etc. However, the form of the user terminal U1 is not limited to such examples. In another example, the information processing device 1 may be the same computer as the user terminal U1.

[0191] §5 Experimental Examples The following experiment was conducted to verify that the large-scale generative model can generate notification conditions, perform the first analysis, and determine the notification conditions. However, the present invention is not limited to the following experimental example.

[0192] [First Experimental Example] First, we used a large-scale generative model to generate pseudo-data of the target data, assuming POS data from a convenience store.

[0193] Figure 15 shows the prompts given to the large-scale generative model to generate pseudo-data. Figure 16 shows the pseudo-data generated by the large-scale generative model. ,ChatGPT(GPT4o-mini) was used.

[0194] (Experimental Example 1-1) In experimental example 1-1, pseudo-data, inputs assuming condition-specified data, and instructions to generate program code were given to a large-scale generative model to generate program code for determining whether notification conditions were satisfied.

[0195] Figure 17 shows the prompts (excluding pseudo data) given to the large-scale generative model in Experimental Example 1-1. Figure 18 shows the program code generated by the large-scale generative model in Experimental Example 1-1. In Experimental Example 1-1, in response to the instruction "sudden increase in beer sales," program code was obtained that executes notification processing in response to a sales increase of 50% or more. These results show that it is possible to generate notification conditions by providing condition specification data as reference data. Furthermore, it was found that as long as a certain degree of notification conditions are given, program code can be generated even if detailed conditions such as thresholds are not given.

[0196] (Experimental Example 1-2) In the first and second experimental examples, we generated metadata to describe pseudo-data. By providing the generated metadata, we had a large-scale generative model propose analysis methods and triggers (notification conditions).

[0197] Figure 19 shows the prompts and metadata provided to the large-scale generative model in Experimental Example 1-2. Figure 20 shows the proposal results obtained from the large-scale generative model in Experimental Example 1-2. In Experimental Example 1-2, the trigger (notification condition) for sales by genre was proposed as "when sales decrease by 20% or more compared to the previous day or the weekly average." This result shows that it is possible to generate notification conditions by providing metadata as reference data. Furthermore, based on the results of Experimental Example 1-1 above, it was inferred that it is also possible to generate program code from metadata.

[0198] (Experimental Examples 1-3) In Experimental Examples 1-3, pseudo data was given to the large-scale generative model to generate notification conditions and program code for determining whether the notification conditions are satisfied.

[0199] Figure 21 shows the prompts (excluding pseudo data) provided to the large-scale generative model in Experimental Example 1-3. Figure 22 shows the results of generating notification conditions obtained from the large-scale generative model in Experimental Example 1-3. Figure 23 shows the results of generating program code obtained from the large-scale generative model in Experimental Example 1-3. In Experimental Example 1-3, notification conditions and program code regarding sales of beer and soft drinks were generated. These results show that it is possible to generate notification conditions by providing at least a portion of the target data as reference data. It also shows that it is possible to generate program code for determining whether notification conditions are satisfied by providing at least a portion of the target data as reference data.

[0200] (Experimental Examples 1-4) In Experiments 1-4, we provided pseudo-data and sample notification conditions to a large-scale generative model to generate notification conditions.

[0201] Figure 24 shows the prompts (excluding pseudo data) given to the large-scale generative model in Experimental Example 1-4. Figure 25 shows the results of generating notification conditions obtained from the large-scale generative model in Experimental Example 1-4. In Experimental Example 1-4, new notification conditions other than the samples could be generated. From these results, it was found that new notification conditions can be generated by providing sample notification conditions as reference data. Furthermore, based on the results of Experimental Example 1-1 above, it was inferred that program code can also be generated from sample notification conditions.

[0202] (Experimental Examples 1-5) In experimental examples 1-5, a large-scale generative model was made to generate notification conditions by giving instructions to search for media data, such as news, and instructions to generate notification conditions from the searched media data.

[0203] Figure 26 shows the prompts given to the large-scale generative model in Experimental Example 1-5. Figure 27 shows the results of generating notification conditions obtained from the large-scale generative model in Experimental Example 1-5. In Experimental Example 1-5, notification conditions were generated for beer sales by category, craft beer sales, low-alcohol / non-alcohol sales, and sales of new beer-related products. These results demonstrate that notification conditions can be generated by providing media data as reference data. Furthermore, based on the results of Experimental Example 1-1 above, it was inferred that program code can also be generated from media data.

[0204] (summary) According to the above Experimental Examples 1-1 to 1-5, it was verified that it is possible to generate notification conditions and their program codes by providing reference data to a large-scale generative model. Furthermore, notification conditions can be an example of task execution conditions. Therefore, it was inferred that it is also possible to generate task execution conditions and their program codes using a similar method.

[0205] [Second Experimental Example] (Experimental Example 2-1) In the 2-1 experimental example, condition instructions and pseudo-data were given to the large-scale generative model to perform an analysis (first analysis) and determine whether the notification conditions were satisfied.

[0206] Figure 28 shows the prompts (excluding pseudo data) given to the large-scale generative model in Experimental Example 2-1. Figure 29 shows the results of determining whether or not the notification conditions were satisfied, obtained from the large-scale generative model in Experimental Example 2-1. In Experimental Example 2-1, it was possible to obtain a determination result as to whether or not to notify (satisfaction of the notification conditions) for each condition given as a condition instruction. These results demonstrate that the large-scale generative model can be used to perform analysis and determine whether or not the notification conditions are satisfied.

[0207] (Experimental Example 2-2) In the second experimental example, instructions were given to the large-scale generative model to search for media data, assuming news and social media information. In addition, instructions and pseudo-data were given to determine whether the notification conditions were met based on the searched media data, causing the large-scale generative model to perform an analysis (first analysis) and determine whether the notification conditions were met.

[0208] Figure 30 shows the prompts (excluding pseudo data) given to the large-scale generative model in Experimental Example 2-2. Figure 31 shows the results of determining whether the notification conditions were satisfied, obtained from the large-scale generative model in Experimental Example 2-2. In Experimental Example 2-2, it was possible to obtain a determination result of whether the notification conditions were satisfied after reflecting the searched media data. These results show that by having the large-scale generative model refer to media data when determining whether the notification conditions are satisfied, it is possible to expect appropriate determinations of whether the notification conditions are satisfied in accordance with social trends.

[0209] (summary) According to Experimental Examples 2-1 and 2-2, it was possible to verify that a large-scale generative model can perform analysis (generation of analysis results) and determine whether notification conditions are satisfied. It was also inferred that it is possible to have a large-scale generative model perform only analysis (generation of analysis results). Furthermore, it was inferred that it is possible to have a large-scale generative model determine whether notification conditions are satisfied by providing analysis results. [Explanation of symbols]

[0210] 1...information processing device, 11...control unit, 12...storage unit, 13...communication module, 81...program, 91...storage medium, 2...Target data, 30...(first analysis) result, 35...(second analysis) result, 4...Notification conditions, 5...Notification data, U0: User terminal

Claims

1. Monitoring whether the results of the first analysis of the continuously collected subject data meet the notification conditions; and notifying a user terminal of notification data indicating a result of a second analysis of the target data when the result of the first analysis satisfies the notification condition as a result of monitoring; a controller configured to execute monitoring whether the result of the first analysis satisfies the notification condition includes causing a second large-scale generative model to perform the first analysis by providing at least a portion of the target data as input, and determining whether the result of the first analysis satisfies the notification condition; Determining whether the result of the first analysis satisfies the notification condition comprises having the second large-scale generative model determine whether the result of the first analysis satisfies the notification condition by referring to media data on social trends related to the target data. Information processing device.

2. Monitoring whether the results of a first analysis of the continuously collected target data satisfy notification conditions; and notifying a user terminal of notification data indicating a result of a second analysis of the target data when the result of the first analysis satisfies the notification condition as a result of monitoring; a controller configured to execute monitoring whether the result of the first analysis satisfies the notification condition includes determining whether the result of the first analysis satisfies the notification condition by analyzing explanatory data generated from the result of the first analysis, the explanatory data explaining the result of the first analysis; Information processing device.

3. At least a portion of the notification conditions are generated by a first large-scale generative model from reference data related to the notification conditions; The reference data includes condition specification data input by a user in natural language. The information processing device according to claim 2 .

4. The notification condition includes a change in the trend of the target data as a result of the first analysis, the target data includes numerical data relating to each of a plurality of items; the first analysis includes calculating a ranking of each of the plurality of items from the numerical data; The change in the trend is evaluated according to the degree of change in the calculated ranking. The information processing device according to claim 2 .

5. Monitoring whether the results of a first analysis of the continuously collected target data satisfy notification conditions; and notifying a user terminal of notification data indicating a result of a second analysis of the target data when the result of the first analysis satisfies the notification condition as a result of monitoring; a controller configured to execute at least a portion of the notification conditions are generated by a first large-scale generative model from reference data related to the notification conditions; The reference data includes condition specification data input by a user in natural language. Information processing device.

6. the reference data further includes metadata describing the target data; The information processing device according to claim 5 .

7. the reference data further includes at least a portion of the target data; The information processing device according to claim 5 .

8. The reference data further includes a sample of the notification condition. The information processing device according to claim 5 .

9. The reference data further includes media data related to social trends related to the target data. The information processing device according to claim 5 .

10. The notification condition includes a change in the trend of the target data as a result of the first analysis, the target data includes numerical data relating to each of a plurality of items; the first analysis includes calculating a ranking of each of the plurality of items from the numerical data; The change in the trend is evaluated according to the degree of change in the calculated ranking. The information processing device according to claim 5 .

11. Monitoring whether the results of a first analysis of the continuously collected target data satisfy notification conditions; and notifying a user terminal of notification data indicating a result of a second analysis of the target data when the result of the first analysis satisfies the notification condition as a result of monitoring; a controller configured to execute the notification condition includes a change in the trend of the target data as a result of the first analysis; the target data includes numerical data relating to each of a plurality of items; the first analysis includes calculating a ranking of each of the plurality of items from the numerical data; The change in the trend is evaluated according to the degree of change in the calculated ranking. Information processing device.

12. The first analysis further includes calculating statistics of the target data; The change in the trend is evaluated further according to the calculated statistical quantity. The information processing device according to claim 11.

13. At least a portion of the notification conditions are generated by a first large-scale generative model from reference data related to the notification conditions; the reference data includes condition specification data input by a user in natural language; monitoring whether the result of the first analysis satisfies the notification condition includes determining whether the result of the first analysis satisfies the notification condition by analyzing explanatory data generated from the result of the first analysis, the explanatory data explaining the result of the first analysis; The information processing device according to claim 11.

14. the notification data includes search information regarding a search result obtained by providing a search engine with a word included in at least one of the result of the second analysis, the notification condition, and the explanatory data of the result of the second analysis, the explanatory data of the results of the second analysis is generated from the results of the second analysis and is configured to explain the results of the second analysis. The information processing device according to claim 1 .

15. The control unit monitoring whether a result of a third analysis of the continuously collected target data satisfies a task execution condition; and performing a predetermined task in response to a result of the third analysis satisfying the task execution condition as a result of monitoring; and further configured to perform The information processing device according to claim 1 .

16. 1. A computer-implemented information processing method, comprising: Monitoring whether the results of the first analysis of the continuously collected subject data meet the notification conditions; and notifying a user terminal of notification data indicating a result of a second analysis of the target data when the result of the first analysis satisfies the notification condition as a result of monitoring; Including, monitoring whether the result of the first analysis satisfies the notification condition includes causing a second large-scale generative model to perform the first analysis by providing at least a portion of the target data as input, and determining whether the result of the first analysis satisfies the notification condition; Determining whether the result of the first analysis satisfies the notification condition comprises having the second large-scale generative model determine whether the result of the first analysis satisfies the notification condition by referring to media data on social trends related to the target data. Information processing methods.

17. An information processing method implemented by a computer, comprising: Monitoring whether the results of the first analysis of the continuously collected subject data meet the notification conditions; and notifying a user terminal of notification data indicating a result of a second analysis of the target data when the result of the first analysis satisfies the notification condition as a result of monitoring; Including, monitoring whether the result of the first analysis satisfies the notification condition includes determining whether the result of the first analysis satisfies the notification condition by analyzing explanatory data generated from the result of the first analysis, the explanatory data explaining the result of the first analysis; Information processing methods.

18. An information processing method implemented by a computer, comprising: Monitoring whether the results of the first analysis of the continuously collected subject data meet the notification conditions; and notifying a user terminal of notification data indicating a result of a second analysis of the target data when the result of the first analysis satisfies the notification condition as a result of monitoring; Including, at least a portion of the notification conditions are generated by a first large-scale generative model from reference data related to the notification conditions; The reference data includes condition specification data input by a user in natural language. Information processing methods.

19. An information processing method implemented by a computer, comprising: Monitoring whether the results of the first analysis of the continuously collected subject data meet the notification conditions; and notifying a user terminal of notification data indicating a result of a second analysis of the target data when the result of the first analysis satisfies the notification condition as a result of monitoring; Including, the notification condition includes a change in the trend of the target data as a result of the first analysis; the target data includes numerical data relating to each of a plurality of items; the first analysis includes calculating a ranking of each of the plurality of items from the numerical data; The change in the trend is evaluated according to the degree of change in the calculated ranking. Information processing methods.

20. A program for causing a computer to execute an information processing method, The information processing method includes: Monitoring whether the results of the first analysis of the continuously collected subject data meet the notification conditions; and notifying a user terminal of notification data indicating a result of a second analysis of the target data when the result of the first analysis satisfies the notification condition as a result of monitoring; Including, monitoring whether the result of the first analysis satisfies the notification condition includes causing a second large-scale generative model to perform the first analysis by providing at least a portion of the target data as input, and determining whether the result of the first analysis satisfies the notification condition; Determining whether the result of the first analysis satisfies the notification condition comprises having the second large-scale generative model determine whether the result of the first analysis satisfies the notification condition by referring to media data on social trends related to the target data. program.

21. A program for causing a computer to execute an information processing method, comprising: The information processing method includes: Monitoring whether the results of the first analysis of the continuously collected subject data meet the notification conditions; and notifying a user terminal of notification data indicating a result of a second analysis of the target data when the result of the first analysis satisfies the notification condition as a result of monitoring; Including, monitoring whether the result of the first analysis satisfies the notification condition includes determining whether the result of the first analysis satisfies the notification condition by analyzing explanatory data generated from the result of the first analysis, the explanatory data explaining the result of the first analysis; program.

22. A program for causing a computer to execute an information processing method, comprising: The information processing method includes: Monitoring whether the results of the first analysis of the continuously collected subject data meet the notification conditions; and notifying a user terminal of notification data indicating a result of a second analysis of the target data when the result of the first analysis satisfies the notification condition as a result of monitoring; Including, at least a portion of the notification conditions are generated by a first large-scale generative model from reference data related to the notification conditions; The reference data includes condition specification data input by a user in natural language. program.

23. A program for causing a computer to execute an information processing method, comprising: The information processing method includes: Monitoring whether the results of the first analysis of the continuously collected subject data meet the notification conditions; and notifying a user terminal of notification data indicating a result of a second analysis of the target data when the result of the first analysis satisfies the notification condition as a result of monitoring; Including, the notification condition includes a change in the trend of the target data as a result of the first analysis; the target data includes numerical data relating to each of a plurality of items; the first analysis includes calculating a ranking of each of the plurality of items from the numerical data; The change in the trend is evaluated according to the degree of change in the calculated ranking. program.

Images