Information processing system, information processing method, and program

Abstract

To provide an information system or the like capable of facilitating an instruction to a AI.SOLUTION: According to an aspect of the present invention, there is provided an information-processing system including one or more processors, wherein the one or more processors are configured to execute, in a reflection step, a reflection process for reflecting a rule-based action in accordance with a predetermined rule in a AI, the rule operation includes an entry field displaying operation for displaying an entry field of an instruction in a natural language, an option generating operation for generating an option of the instruction, an option displaying operation for displaying the generated option, a response generating operation for generating a response according to the instruction entered in the entry field or the instruction indicated by the selected option, and a response displaying operation for displaying the generated response, and the reflection process includes a providing process for providing rule information indicating the rule and a sample of the option to the AI.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

[0002] Patent Document 1 discloses a technology for providing a customized artificial intelligence model by acquiring selection information for identifying multiple selected models selected from multiple artificial intelligence models each with different processing content, creating an artificial intelligence model set including multiple selected models corresponding to the acquired selection information, and outputting the created artificial intelligence model set. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-109159 Summary of the Invention [Problem to be solved by the invention]

[0004] Customized AI can respond to various commands from the user, but it is difficult for the user to give the appropriate commands to get the response they want.

[0005] In view of the above circumstances, the present invention provides an information processing system and the like that can easily give instructions to AI. [Means for solving the problem]

[0006] According to one aspect of the present invention, there is provided an information processing system having one or more processors, wherein the processor(s) execute a reflection process in a reflection step to reflect rule behavior in accordance with predetermined rules in an AI having the function of responding to instructions in natural language, the rule behavior including an input field display action that displays an input field for instructions in natural language, an option generation action that generates option options for the instructions, an option display action that displays the generated option options, a response generation action that generates a response according to the instruction entered in the input field or the instruction indicated by the selected option, and a response display action that displays the generated response, and the reflection process includes a provision process that provides rule information indicating the rule and sample options to the AI.

[0007] According to this embodiment, it is possible to easily give instructions to the AI. [Brief explanation of the drawings]

[0008] [Figure 1] FIG. 1 is a diagram illustrating an example of the overall configuration of a custom AI processing system 1. [Figure 2] FIG. 2 is a diagram illustrating an example of the hardware configuration of a custom AI server 10. [Figure 3] FIG. 2 is a diagram illustrating an example of a hardware configuration of a user terminal 20. [Figure 4] FIG. 10 is an activity diagram illustrating an example of a creation process. [Figure 5] FIG. 10 is a diagram illustrating an example of a custom AI creation screen. [Figure 6] FIG. 10 is a diagram showing an example of an input "role." [Figure 7] FIG. 10 is a diagram showing an example of input "essential work." [Figure 8] FIG. 10 is a diagram showing an example of saved custom information and custom AI. [Figure 9] FIG. 10 is a diagram illustrating an example of a custom AI editing screen. [Figure 10] FIG. 10 is an activity diagram illustrating an example of operation processing. [Figure 11] FIG. 10 is a diagram showing an example of a custom AI selection screen. [Figure 12] FIG. 10 is a diagram showing an example of an instruction input screen for a custom AI. [Figure 13] FIG. 10 is a diagram showing an example of a displayed instruction input screen. [Figure 14] FIG. 10 is a diagram showing an example of a continuation of the instruction input screen. [Figure 15] FIG. 10 is a diagram illustrating an example of a sample table. [Figure 16] FIG. 10 is a diagram illustrating another example of the sample table. [Figure 17] FIG. 10 is a diagram showing an example of displayed parameter options. [Figure 18] FIG. 10 is a diagram illustrating an example of a number table. [Figure 19] FIG. 10 is a diagram illustrating another example of the number table. [Figure 20] FIG. 10 is a diagram illustrating an example of an option table. DETAILED DESCRIPTION OF THE INVENTION

[0009] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below with reference to the accompanying drawings. Various features shown in the following embodiments can be combined with each other.

[0010] Incidentally, the program for realizing the software appearing in one embodiment may be provided as a non-transitory computer-readable medium, or may be provided so that it can be downloaded from an external server, or may be provided so that the program is started on an external computer and its functions are realized on a client terminal (so-called cloud computing).

[0011] Furthermore, various information processing according to an embodiment may realize input and output corresponding to the input. Here, the form of information referenced in such information processing (hereinafter referred to as reference information) is not limited as long as an output is obtained as a result of the input. The reference information may be, for example, rule-based information such as a database, a lookup table, or a predetermined function (including a decision formula such as a regression formula constructed using a statistical method), a trained model that has previously trained the correlation between input and output, or a large-scale language model that can output a desired result by inputting a prompt.

[0012] In one embodiment, a "unit" may include, for example, a combination of hardware resources implemented by a circuit in the broad sense and software information processing that can be specifically realized by these hardware resources. In one embodiment, various information is handled, and this information is represented, for example, by physical values ​​of signal values ​​representing voltage and current, high and low signal values ​​as a binary bit set consisting of 0 or 1, or quantum superposition (so-called quantum bits), and communication and calculations can be performed on a circuit in the broad sense.

[0013] Furthermore, a circuit in the broad sense is a circuit realized by at least an appropriate combination of a circuit, circuitry, processor, memory, etc. The processor may be a general-purpose processor or a dedicated circuit. That is, it includes an application specific integrated circuit (ASIC), a programmable logic device (e.g., a simple programmable logic device (SPLD), a complex programmable logic device (CPLD), and a field programmable gate array (FPGA)), etc.

[0014] <Embodiment> 1. System Configuration The system configuration according to the embodiment will be described below. Figure 1 is a diagram showing an example of the overall configuration of a custom AI processing system 1. Figure 1 shows an overview of each device included in the custom AI processing system 1 and the users who use those devices. Each overview will be explained as needed, with reference to other figures.

[0015] The custom AI processing system 1 is an information processing system that executes information processing such as creation processing for creating custom AI (Artificial Intelligence) and operation processing for operating the custom AI. Custom AI is AI whose behavior has been customized by a user. Users of the custom AI processing system 1 include system operators and service users. The system operator is, for example, a user who performs the work of creating a custom AI, and the service user is, for example, a user who uses a service based on the created custom AI.

[0016] The custom AI processing system 1 comprises a communication line 2, a custom AI server 10, a user terminal 20, an operator terminal 30, and a service providing server 40. The communication line 2 is not particularly limited, but may be configured, for example, by the Internet. The communication line 2 may also include a local area network, a mobile communication network, a VPN (Virtual Private Network), etc. The communication line 2 mediates the exchange of data between devices connected to the communication line. In the example of FIG. 1, the custom AI server 10 and the service providing server 40 are connected to the communication line 2 by wire, and the user terminal 20 and the operator terminal 30 are connected wirelessly. Note that the connection of each device to the communication line 2 may be wired or wireless.

[0017] The custom AI server 10 is an information processing device that executes information processing such as the creation processing and operation processing described above. The custom AI server 10 includes an AI module 100. The AI ​​module 100 is a module that is adjusted to realize predetermined functions using AI technology. The predetermined functions include, for example, a response function that responds in natural language to questions from service users in natural language. The predetermined functions also include multiple types of analysis functions, which will be explained later. The AI ​​module 100 is the AI ​​that serves as the basis for customization, and is hereinafter also referred to as the "base AI."

[0018] The service providing server 40 is an information processing device that executes information processing to provide services used by service users. In the example of FIG. 1, the service providing server 40 provides financial accounting services and stores data related to the financial accounting of service users. The AI ​​module 100 is adjusted to realize, as a predetermined function, the function of aggregating the data related to the financial accounting of service users stored by the service providing server 40.

[0019] The user terminal 20 is a terminal whose user is a user of the service provided by the custom AI processing system 1, and is, for example, a personal computer, smartphone, or tablet terminal. The user terminal 20 displays a service usage screen and accepts operations by service users. The operator terminal 30 is a terminal whose user is a system operator who operates the custom AI processing system 1, and is, for example, a personal computer. The operator terminal 30 displays an operation screen for operating the system and accepts operations by the system operator.

[0020] The custom AI server 10 performs display processing to display images on the user terminal 20 and the operator terminal 30, and authentication processing to authenticate the service users and system operators who use the user terminal 20 and the operator terminal 30.

[0021] The custom AI server 10 performs processes such as generating and transmitting HTML (Hyper Text Markup Language) files as display processing, and uses browser functions to display web pages showing system screens on the user terminal 20 and the operator terminal 30. Note that the user terminal 20 and the operator terminal 30 may install an application program for using the custom AI processing system 1, and the custom AI server 10 may perform processes such as generating and transmitting display data in that application as display processing. The custom AI server 10 controls the displays of the user terminal 20 and the operator terminal 30 by performing these display processes.

[0022] The custom AI server 10 stores authentication information (user ID, password, etc.) for authenticating users who use the custom AI processing system 1, such as service users and system operators, and authenticates users who input the authentication information. By authenticating users, the custom AI server 10 can restrict access to data and assign identification information to data entered by users to make it identifiable.

[0023] 2. Hardware Configuration The hardware configuration according to the first embodiment will be described below. 2 is a diagram showing an example of the hardware configuration of custom AI server 10. Custom AI server 10 includes a control unit 11, a storage unit 12, a communication unit 13, and a bus 14. Bus 14 electrically connects the various units included in custom AI server 10.

[0024] (Control unit 11) The control unit 11 has at least one processor. The at least one processor may be configured by, for example, a central processing unit (CPU), a micro processing unit (MPU), a graphics processing unit (GPU), one or more integrated circuits, one or more discrete circuits, or a combination thereof (not shown).

[0025] The control unit 11 is a computer that implements various functions related to the custom AI processing system 1 by reading out predetermined programs stored in the storage unit 12. In other words, information processing by software stored in the storage unit 12 is specifically implemented by the control unit 11, which is an example of hardware, and can be executed as each functional unit included in the control unit 11. Note that the control unit 11 is not limited to being single, and multiple control units 11 may be provided for each function. A combination of these may also be used.

[0026] (Storage unit 12) The memory unit 12 stores various pieces of information defined above. This can be implemented, for example, as a storage device such as a solid state drive (SSD) or hard disk drive (HDD) that stores various programs and the like related to the custom AI processing system 1 executed by the control unit 11, or as a memory such as random access memory (RAM) that stores temporarily required information (arguments, arrays, etc.) related to the program operations. The memory unit 12 stores various programs, variables, etc. related to the custom AI processing system 1 executed by the control unit 11.

[0027] (Communications Department 13) The communication unit 13 is configured with a communication module. The communication module may be a wireless communication module conforming to standards such as IEEE802.11a / b / g / n / ac / ax, LTE, 5G, or 6G, or may be a wired communication module conforming to standards such as IEEE802.3. The communication unit 13 is configured to be able to transmit various electrical signals from the custom AI server 10 to external components. The communication unit 13 is also configured to be able to receive various electrical signals from external components to the custom AI server 10. More preferably, the communication unit 13 has a network communication function, which allows various information to be communicated between the custom AI server 10 and external devices via the communication line 2.

[0028] Fig. 3 is a diagram showing an example of the hardware configuration of user terminal 20. User terminal 20 includes control unit 21, memory unit 22, communication unit 23, input unit 24, output unit 25, and bus 26. Bus 26 electrically connects the various units included in user terminal 20. Control unit 21, memory unit 22, and communication unit 23 are similar hardware to control unit 11, memory unit 12, and communication unit 13 shown in Fig. 2, although their specifications, models, etc. may differ.

[0029] (Input unit 24) The input unit 24 has keys, buttons, a touch screen, a mouse, etc., and receives input from the user. The input unit 24 may also have a microphone and have the function of receiving voice input from the user.

[0030] (Output section 25) The output unit 25 has a display, a speaker, etc., and displays visual information generated in a manner that is visible to the user, such as a screen, an image, an icon, or text, on the display surface of the display, and outputs sound including voice.

[0031] The service providing server 40 shown in Figure 2 has the same hardware configuration as the custom AI server 10. In the service providing server 40, only the control unit 41 will be described with a different reference number from the control unit 11 of the custom AI server 10. The operator terminal 30 shown in Figure 3 has the same hardware configuration as the user terminal 20. In the operator terminal 30, only the control unit 31 will be described with a different reference number from the control unit 21 of the user terminal 20.

[0032] 3. Information Processing The information processing according to the embodiment will be described below. In the following description, the custom AI server 10, user terminal 20, operator terminal 30, and service provider server 40 are described as the main entities performing each information process, but this information processing is executed by at least one processor included in the custom AI processing system 1, i.e., the processor included in the control units 11, 21, 31, and 41 of each device. The custom AI processing system 1 executes the custom AI creation process and operation process described above.

[0033] 4 is an activity diagram showing an example of the creation process. The creation process is executed when the system operator performs an AI customization operation on the operator terminal 30. First, the operator terminal 30 displays a custom AI creation screen for creating a custom AI (activity A11).

[0034] FIG. 5 is a diagram showing an example of a custom AI creation screen. The custom AI creation screen C1 shown in FIG. 5 displays the character string "Please enter custom information," custom information input fields D11 to D21, and a customization execution button B11. The custom information is information for customizing the AI ​​module 100, which is the base AI. The input fields D11 to D21 are used to input the following custom information: "Custom AI name," "Role," "Basic rules," "Constraints," "Required tasks," "API rules," "Analysis rules," "Calculation rules," "Display rules," "External data rules," and "Graph creation rules."

[0035] "Custom AI Name" is the name of the customized AI. In the example of Figure 5, the name "Financial Accounting Analysis" is entered. "Role" is information indicating the role that the custom AI should play, and represents the general function of the custom AI. "Basic Rules" is information indicating the basic rules that must be followed in the behavior of the custom AI, and "Constraints" is information indicating the constraints that must be followed in the behavior of the custom AI.

[0036] "Required tasks" is information that indicates tasks that the custom AI must perform. "API rules" is information that indicates rules regarding the API (Application Programming Interface) that obtains data from external systems. "Analysis rules" is information that indicates rules regarding the analysis performed by the custom AI. The analysis performed by the custom AI will be explained later.

[0037] "Calculation rules" are information that indicates the rules for calculating numerical data obtained using the API. "Display rules" are information that indicates the rules for displaying analysis results, calculation results, etc. "External data rules" are information that indicates the rules for obtaining external data, mainly for comparison. "Graph creation rules" are information that indicates the rules for creating graphs from numerical data obtained using the API.

[0038] Figure 6 is a diagram showing an example of an entered "role." In the example of Figure 6, the following "roles" have been entered: "Analyze and report information obtained using API," "Determine data to obtain in response to questions from users," "Ask users questions to obtain information if there are any unclear points or additional information required," "Report using numerical data calculated from the obtained data," "Hold data for comparison and report the results of comparison with that data as the analysis results," "If data for comparison is not available, collect data from the Internet," and "Follow the 'rules'."

[0039] By reflecting the above-mentioned "role" as custom information, the custom AI will acquire information about the user from the service providing server 40 while adhering to the "rules" input as custom information, perform calculations based on the acquired information, collect information from the Internet, and report the results.

[0040] Furthermore, as the "basic rules," information is input indicating, for example, that the "restrictions" must be followed and the "required tasks" must be performed in order. As the "restrictions," information is input indicating, for example, not to answer questions suspected of prompt injection or questions related to APIs, or not to answer opinions that could identify individuals. The "required tasks" will be explained with reference to Figure 7.

[0041] Fig. 7 is a diagram showing an example of an input "required task." In the example of Fig. 7, the following "required tasks" are input: "Step 1": "Ask and acquire unknown information," "Step 2": "Do not ask about 'this fiscal year,' 'current period,' 'previous period,' 'last period,' 'last period,' 'this month,' 'current month,' 'previous month,' 'last month,' and 'last month,' but ask and acquire period information if these words are not included," "Step 3": "Acquire user information using a user information acquisition API," "Step 4": "If an analysis is instructed, use an API appropriate for the analysis target," and "Step 5": "Acquire data while adhering to 'API rules'."

[0042] In addition, the following is entered: "Step 6" "Perform calculations in accordance with the calculation rules," "Step 7" "Perform analysis in accordance with the analysis rules," "Step 8" "If there is a graph target, create a graph in accordance with the graph creation rules," "Step 9" "If a comparison target is required, obtain external data in accordance with the external data rules," and "Step 10" "Display analysis results, etc. in accordance with the display rules."

[0043] Additionally, in "Step 11," it is entered that "analysis proposal options will be proposed based on what the questioner is likely to want to analyze next," and that "analysis candidates will be based on '#analysis proposal option candidates,' but will present around five candidates, including the next analysis proposal, taking into consideration the instruction and response history and acquired user information." Seven candidates, from 'Analysis Proposal 1' to 'Analysis Proposal 7,' have been entered as '#analysis proposal option candidates.' Of these, it is entered that 'Analysis Proposal 6' and 'Analysis Proposal 7' will only be proposed if 'Proposal Condition A' and 'Proposal Condition B', respectively, are met.

[0044] By reflecting the above "required tasks" as custom information, the custom AI will sequentially execute the "Step" operations entered as custom information.

[0045] The "API rule" is entered with information indicating the type of API and the definition of the terms used when obtaining information with that type of API. For example, the terms "this fiscal year," "current period," "previous period," "last period," and "last period" are defined as specifying the fiscal year calculated based on the current date and the accounting period obtained by the API, while the terms "this month," "current month," "previous month," and "last month" are defined as specifying the month calculated based on the current date.

[0046] Regarding "analysis rules," we will first explain the analytical functions of the custom AI. The custom AI has the same multiple analytical functions as the base AI described above, and can perform various types of analysis. Examples of various analyses include analysis of profit and loss status, comparison of profit and loss status and analysis of causes of changes and areas for improvement, analysis of financial soundness and safety, and comparative analysis of financial soundness and stability. "Analysis rules" indicate the rules for these various types of analysis. For example, information indicating the definition of important items in the analysis (such as "industry," "business type," or "company size") or the definition of the range of data to be used in the analysis is input as "analysis rules."

[0047] The "calculation rules" include information indicating combinations of data that should be used in calculations (such as debits and credits within the same financial statement item or account item) or combinations of data that should not be used in calculations (such as amounts of different financial statement items or account item).

[0048] The "display rules" contain information indicating points to note when displaying data. Specifically, information indicating points to note such as whether to leave the sign as is when displaying numerical values ​​or convert it to absolute values, and how to label numerical values ​​(for example, "cumulative total" when analyzing and displaying data obtained from an income statement, and "balance" when analyzing and displaying data obtained from a balance sheet, etc.) is entered.

[0049] The "external data rules" include information indicating the conditions for permitting the acquisition of information from the Internet, the conditions for permitting the reading of files attached by a user, etc. The "graph creation rules" include information indicating the type of graph and the type of legend, etc.

[0050] The operator terminal 30 accepts input of custom information into the custom AI creation screen C1 (activity A12). Next, the operator terminal 30 accepts an operation on the customization execution button B11 as an operation to create and execute a custom AI (activity A13), and sends the input custom information to the custom AI server 10. The custom AI server 10 acquires the sent custom information (activity A21).

[0051] When the custom AI server 10 acquires the custom information, it reads, for example, the AI ​​module 100 described above as a base AI that serves as the basis for the custom AI (activity A22). Next, the custom AI server 10 executes a reflection process that reflects the acquired custom information in the read base AI (activity A23). The custom AI server 10, for example, generates training data that sets the behavior indicated by the custom information as the correct answer, and causes the AI ​​module 100 to perform additional learning using the generated training data, thereby reflecting the custom information in the AI ​​module 100. Note that the method of reflecting the custom information is not limited to this, and any well-known technology may be used. When the reflection process is complete, the custom AI server 10 saves the custom information and the AI ​​module 100 that reflects the custom information, i.e., the custom AI (activity A24).

[0052] FIG. 8 is a diagram showing an example of saved custom information and custom AI. In the example of FIG. 8, custom AI, custom information, and storage date and time are stored in association with each other in custom AI database DB1. In addition to the information shown in FIG. 8, information such as the creator's user ID or base AI may also be stored. Saved custom AI can be updated by editing the custom information. In this case, the operator terminal 30 displays a custom AI editing screen (activity A31).

[0053] Figure 9 is a diagram showing an example of a custom AI editing screen. The custom AI editing screen C2 shown in Figure 9 displays the text "Please edit your custom information," custom information input fields D11 to D21 similar to those in Figure 5, and a custom update button B21. When the system operator enters the name of a custom AI in input field D11, the custom AI server 10 reads out the custom information stored in association with the custom AI with the entered name (activity A32) and displays it in each input field.

[0054] The operator terminal 30 accepts an operation to edit the custom information in each input field (activity A33), and accepts an operation on the customization update button B21 as an operation to update the custom information (activity A34). When the operator terminal 30 accepts the operation to update, it sends the edited custom information to the custom AI server 10. The custom AI server 10 acquires the sent custom information (activity A41).

[0055] When the custom AI server 10 acquires the custom information, it reads out the custom AI stored in association with the acquired custom information (activity A42). Next, the custom AI server 10 executes a reflection process to reflect the acquired custom information in the read custom AI (activity A43). When the reflection process is complete, the custom AI server 10 overwrites and saves the custom information and the AI ​​module 100 that reflects the custom information, i.e., the custom AI (activity A44).

[0056] The operation process for operating the custom AI created and updated as described above will now be described. FIG. 10 is an activity diagram showing an example of operation processing. FIG. 10 includes information processing executed by the custom AI server 10 using the custom AI function. This information processing is executed by the control unit 11 of the custom AI server 10, but for simplicity, the explanation will be centered on the custom AI. First, the operator terminal 30 displays a custom AI selection screen for selecting the custom AI to operate (activity A51).

[0057] Figure 11 is a diagram showing an example of a custom AI selection screen. The custom AI selection screen C3 shown in Figure 11 displays the character string "Please select a custom AI" and a list D31 of custom AI names. The operator terminal 30 accepts the operation of selecting a name displayed in the list D31 as a custom AI selection operation (activity A52), and notifies the custom AI server 10 of the name of the selected custom AI.

[0058] The custom AI server 10 reads the custom AI with the notified name from the custom AI database DB1 (activity A53). Using the custom AI function, the custom AI server 10 generates an AI input screen for accepting input of instructions to the read custom AI (activity A54), and displays the AI ​​input screen on the operator terminal 30 (activity A55).

[0059] FIG. 12 is a diagram showing an example of an instruction input screen for a custom AI. The instruction input screen C4 shown in FIG. 12 displays a character string "Please enter instructions for the custom AI." and an input field D41 for instructions to the custom AI. In the example of FIG. 12, an instruction "Please analyze this month's profit and loss situation and report it" is entered in the input field D41. When, for example, the execute key is operated after the character string shown in FIG. 12 is entered, the operator terminal 30 accepts it as input of the instruction indicated by the character string (activity A56) and transmits instruction information indicating the entered instruction to the custom AI server 10.

[0060] The custom AI server 10 inputs the received instruction information into the custom AI (activity A61). The custom AI generates a response to the instruction indicated by the input instruction information. Specifically, the custom AI performs the task of generating a response in accordance with the "Steps" shown in the "Required Tasks" in Figure 7. First, if the information required for the instructed analysis is not available, the custom AI generates a response that asks a question about the unknown information (Steps 1 and 2).

[0061] If the custom AI has all the information necessary for analysis, it uses the API to obtain the user information necessary for analysis (Steps 3, 4, and 5). The custom AI first requests user information from the service providing server 40 using the API (activity A62). The service providing server 40 reads the requested user information (activity A63) and sends it to the custom AI server 10. The custom AI then obtains the user information that has been sent (activity A64).

[0062] Next, the custom AI calculates the values ​​necessary for the analysis based on the information included in the instructions and the acquired user information (Step 6).The custom AI then performs the instructed analysis based on the acquired information and calculation results (Step 7).If the custom AI determines that it would be better to generate a graph showing the analysis results, it generates the graph (Step 8), and if it determines that external data is necessary for comparison, it acquires the external data (Step 9).

[0063] The custom AI performs the tasks in Steps 3 to 9 while adhering to each rule (API rule, calculation rule, analysis rule, graph creation rule, external data rule). The custom AI generates a response including the analysis results generated by the above tasks as a response to the above-mentioned instruction from the user (activity A65). Next, the custom AI generates analysis proposal options (activity A66). The custom AI generates approximately five analysis proposal options based on the "analysis proposal option candidates" shown in the "Required Tasks" in Figure 7, the history of instructions and responses to date, and the acquired user information.

[0064] For example, the custom AI initially generates most of the options by selecting them from the analysis proposal option candidates, but as those options become selected, it begins to generate new options. Furthermore, the custom AI generates specific options, such as analysis proposals 6 and 7, only when the corresponding proposal conditions (e.g., proposal conditions A and B) are met. The custom AI then generates an instruction input screen including the generated responses and options (activity A64) and displays the generated instruction input screen on the operator terminal 30 (activity A71).

[0065] FIG. 13 is a diagram showing an example of a displayed instruction input screen. FIG. 14 is a diagram showing an example of a continuation of the instruction input screen. The instruction input screen C4 shown in FIGS. 13 and 14 displays a response display field D42 in addition to the input field D41 for instructions to the custom AI shown in FIG. 12. In the display field D42, a communication log E41 with the service providing server 40 using the API and an analysis result E42 generated by the custom AI are displayed as a response. After the analysis result E42, a character string saying "Please let us know what you would like to analyze next or if you have any other questions," and five analysis suggestions generated by the custom AI are displayed as options E43.

[0066] The options E43 include four analysis proposals (analysis proposals 2 to 5) that were "candidate analysis proposal options" and a newly generated proposal. In the example of FIG. 14, the content of analysis proposal 1 was similar to the content of the analysis specified in input field D41, and therefore was determined not to be "the item that the user would like to analyze next." Note that, as described above, the custom AI originally has a response function that generates responses in natural language, and therefore can generate and display analysis proposals other than the "candidate analysis proposal options." In this way, the custom AI generates options based on past instructions and responses. Following the display field D42, the custom AI server 10 displays an input field D43 for entering the next instruction.

[0067] The operator terminal 30 accepts input of an instruction by selection from the displayed options or an instruction by a character string (activity A72). Note that an option is selected, for example, by inputting the option number in the input field D43 (for example, by inputting "2" to select the second option). The operator terminal 30 transmits instruction information indicating the input instruction to the custom AI server 10. Thereafter, the operations from activity A61 (input of instruction information) onwards are repeated, and the analysis required by the user is performed.

[0068] As described above, the custom AI server 10 functions as an example of a reflection unit that executes reflection processing. Reflection processing is processing for reflecting regular behavior in accordance with predetermined rules in an AI (the base AI described above) that has the function of responding to instructions in natural language.

[0069] The rule actions include an input field display action, an option generation action, an option display action, a response generation action, and a response display action. The input field display action is an action of displaying an input field for instructions in natural language. The input field display action includes, for example, an action of displaying the input field D41 shown in FIG. 12 and the input field D43 shown in FIG. 14. The option generation action is an action of generating options for instructions in natural language. The option generation action includes, for example, an action of generating option E43 shown in FIG. 14.

[0070] The option display operation is an operation of displaying the generated options. The option display operation includes, for example, an operation of displaying option E43 shown in FIG. 14. The response generation operation is an operation of generating a response in response to an instruction by a user. The instruction by the user includes an instruction entered in an input field displayed by the input field display operation and an instruction indicated by an option selected from the options displayed by the option display operation. The response display operation is an operation of displaying a response generated by the response generation operation. The response display operation includes, for example, an operation of displaying the analysis result E42 shown in FIGS. 13 and 14 as a response to an instruction by the user.

[0071] The reflection process also includes a provision process that provides rule information indicating the predetermined rules to the AI. The AI ​​here includes both base AI and custom AI. For example, in A23 of Figure 4, provision processing is performed for the base AI, and in A43 of Figure 4, provision processing is performed for the custom AI.

[0072] The rule information includes, for example, the custom information shown in FIG. 5. For example, "role" is information indicating rules for operations such as obtaining information using an API, analyzing information, creating support, judging data, asking users questions, and collecting data. "Required tasks" is information indicating rules for operations such as asking users questions, obtaining information using an API, calculating numerical values, analyzing information, creating graphs, obtaining external data, displaying analysis results, and presenting options. "Basic rules," "API rules," "analysis rules," "calculation rules," "display rules," "external data rules," and "graph creation rules" are information indicating rules for various operations, and "constraints" is information indicating rules for operation constraints.

[0073] The reflection process also includes a process of providing the AI ​​with sample options described above. The sample options are, for example, the seven "analysis suggestions" shown as "candidate analysis suggestion options" in Figure 7. By providing sample options in this way, options are consistently displayed in the custom AI's response, such as option E43 shown in Figure 14. A user uses the custom AI by repeatedly giving instructions and responding, and the consistently displayed options allow the user to select from the options without having to think of instructions from scratch, making it easier to give instructions to the AI ​​than if sample options were not provided.

[0074] Furthermore, in an AI that performs additional learning, such as a custom AI, repeated additional learning can cause the AI ​​to behave in a way that is not intended by the user who performed the customization (for example, in the above example, behavior that displays options that cause the AI ​​to perform an action different from the analysis).By providing sample options as described above, the custom AI processing system 1 can make it easier to continue generating options that are in line with the intentions of the user who performed the customization, in other words, it is easier to control the direction of the options.

[0075] As described above, the base AI has multiple types of analysis functions (such as a function for analyzing profit and loss situations). The user's instructions specify the information to be analyzed and the type of analysis function to analyze that analysis target. The information to be analyzed is, for example, information for each financial or accounting period (such as the amount of profit and loss for this month or the amount of profit and loss for last month). The types of analysis functions include, for example, analysis of a single analysis target, analysis comparing two or more analysis targets, analysis of the causes of change, or analysis of areas for improvement.

[0076] The rule action also includes an acquisition action that acquires a specified analysis target. The acquisition action is, for example, an action that acquires user information from the service providing server 40 using an API. The response generation action includes an action that generates the results of analyzing the acquired analysis target using a specified type of analysis function as a response to a user instruction. This aspect makes it easier for the user to instruct an analysis than when the analysis results are not used as a response.

[0077] The reflection process also includes a definition process that provides the AI ​​with definitions of the terms used when specifying the analysis target and analysis function. For example, in the example in Figure 5, the AI ​​is provided with definitions of the terms "this fiscal year," "current period," "previous period," "last period," and "last period" (periods specified by calculating the fiscal year based on the current date and the accounting period obtained via API), and definitions of the terms "this month," "current month," "previous month," and "last month" (periods specified by calculating the month based on the current date).

[0078] According to this aspect, since inaccurate terms can cause erroneous analysis, by correctly defining terms, the accuracy of the analysis can be increased compared to when definitions of terms are not provided.

[0079] The reflection process also includes an explanation process that provides the AI ​​with an explanation of how to analyze the analysis target. For example, in the example in Figure 5, definitions of important items in the analysis or the range of data used in the analysis are provided to the AI ​​as an explanation of how to analyze the analysis target. Even an AI with analytical capabilities does not necessarily perform the analysis intended by the user, but by performing the explanation process described above, it becomes easier to obtain the analysis results intended by the user compared to when the explanation process is not performed.

[0080] <Variation: Automatic update of custom information> In the example of Figure 4, the custom information is updated by the user's update execution operation, but it may also be updated automatically. For example, the provision process includes a process of providing the AI ​​(base AI and custom AI) with sample analysis instruction options including the added analysis target when information that can be acquired as an analysis target is added. The custom AI server 10 performs this provision process, for example, using a sample table that associates types of user information with sample analysis options.

[0081] Fig. 15 is a diagram showing an example of a sample table. In the sample table TB1 shown in Fig. 15, sample options such as "Analysis of profit and loss situation," "Comparative analysis of profit and loss situation," and "Analysis of causes of problems in profit and loss situation" are associated with a type of user information called "accounting information." In addition, sample options such as "Analysis of financial soundness and safety," "Comparative analysis of financial soundness and safety," and "Analysis of causes of problems in financial soundness and safety" are associated with a type of user information called "financial information."

[0082] For example, if initially only transaction information was available as user information from the service provider server 40 using an API, custom information containing only sample options related to the analysis of transaction information is provided to the base AI when the custom AI is created. Suppose financial information later becomes available using the API. In this case, the custom AI server 10 determines that new information has been added as an analysis target, and executes a provision process to add sample options associated with "financial information" in sample table TB1 to the custom information as sample analysis instructions containing the added financial information, and to reflect this in the custom AI. This implementation allows the user to issue instructions for analysis of the added analysis target without any effort.

[0083] Furthermore, when a new analysis function is added to the AI, the provision process may include a process of providing the AI ​​with sample analysis instruction options using the added function. Adding an analysis function to an AI is basically performed on the base AI, but it may also be performed on a custom AI. The custom AI server 10 performs this provision process using, for example, a sample table that associates the type of analysis function with sample analysis options.

[0084] Fig. 16 is a diagram showing another example of a sample table. In the sample table TB2 shown in Fig. 16, sample options such as "Analysis of profit and loss situation," "Comparative analysis of profit and loss situation," and "Analysis of causes of problems in profit and loss situation" are associated with the type of analysis function called "Accounting information analysis function." In addition, sample options such as "Analysis of financial soundness and safety," "Comparative analysis of financial soundness and safety," and "Analysis of causes of problems in financial soundness and safety" are associated with the type of analysis function called "Financial information analysis function."

[0085] First, we will explain the case where additional learning is performed on a base AI and an analytical function is added. In this case, even if there is a custom AI that already reflects the custom information, the custom AI server 10 will reflect the custom information again in the base AI to which the analytical function has been added. At that time, the custom AI server 10 adds sample options that are associated with the type of added analytical function in the sample table TB2 to the sample options included in the custom information, and reflects the custom information after the sample addition in the base AI.

[0086] Next, we will explain the case where additional learning is performed on a custom AI and an analytical function is added. In this case, the custom AI server 10 adds a sample option that is associated with the type of added analytical function in the sample table TB2 to the sample options included in the custom information of the custom AI. Then, the custom AI server 10 reflects the custom information after the sample addition in the custom AI, similar to A43 shown in FIG. 4. According to this embodiment, the user can instruct analysis using the added analytical function without any effort.

[0087] <Variation: Analysis parameters> One or more parameters may be used to instruct an analysis. Parameters include, for example, a time period, a range of values, a ranking, an industry, a business type, or a company size. For example, in an "analysis of profit and loss situation," parameters can be set such as which "time period" to target, what "range" of loss and profit amounts for each month to target, what "rank" from the largest amount to target, and which "industry," "business type," or "company size" to target in a comparative analysis.

[0088] In this case, the reflection process may include a candidate providing process that provides parameter candidates to the AI. Specifically, for example, by including "#Parameter Option Candidate" in the custom information, similar to "#Analysis Proposal Option Candidate" shown in FIG. 7, the custom AI server 10 provides the AI ​​with the parameter candidates entered as "#Parameter Option Candidate." Also, for example, suppose the following sentence is added to the "Required Task" shown in FIG. 7: "Propose parameter options based on the parameters you want to use in the analysis selected from the options. The parameter options are based on "#Parameter Option Candidate," but approximately five options will be presented, including suggested parameters to be used, taking into account the instruction / response history and acquired user information."

[0089] In this case, the rule action includes a parameter generation action that generates parameters to be used in the instructions indicated by the selected option, and a parameter display action that displays the generated parameters. A custom AI that has undergone a reflection process that reflects such rule actions displays parameter candidates as follows:

[0090] Fig. 17 is a diagram showing an example of displayed parameter options. On the instruction input screen C4 shown in Fig. 17, following the input field D43 for instructions to the custom AI shown in Fig. 14, a response display field D44 and an input field D45 for the next instruction are displayed. Analysis proposal 4 has been selected and entered from the options shown in Fig. 14 in the input field D43. The display field D44 displays the character string "Please let us know if there are any parameters you would like to use in the analysis," along with five parameter options E44, parameter proposals 1 to 5.

[0091] Option E44 is a proposed parameter, such as a time period, numerical range, or ranking, generated by the custom AI based on the content of the "required tasks" described above. If the user has suitable parameters, they can enter their numbers in input field D45 to perform an analysis using the selected parameters. By providing parameter candidates in this way, parameter options are consistently displayed in the custom AI's response. As a result, users can select from the options without having to think of parameters from scratch, making parameter specification easier than if parameter candidates were not provided to the AI.

[0092] <Variation: Number of options> Regarding the number of options to be displayed, the custom information shown in Fig. 7 indicates a rule action of "present about 5 analysis candidates" as the option display action, but the option display action is not limited to this. For example, the option display action may include an action to limit the number of options to be displayed to a predetermined number or less.

[0093] In this case, the custom information may indicate an option display operation such as "Analysis candidates are always presented in five or fewer numbers." Note that the predetermined number is not limited to "five" and may be any number equal to or greater than one. This embodiment can prevent the user from having too many options, making it difficult to select, compared to when the number of options is not limited.

[0094] The custom AI server 10 may also decrease the predetermined number in accordance with the number of times the options have been displayed. In this case, the custom AI server 10 uses a number table that associates the number of times an option has been displayed with the upper limit of the number of options that can be displayed.

[0095] Figure 18 is a diagram showing an example of a number table. In the number table TB3 shown in Figure 18, when the "number of times an option is displayed" is "0," "1, 2," "3, 4," and "5," the "maximum number of options displayed" is associated with "5," "4," "3," and "2." The custom AI server 10 counts the number of times an option is displayed, such as option E43 shown in Figure 14, and displays the number of options associated with the counted number in the number table TB3.

[0096] The custom AI server 10 may execute a reflection process so that the custom AI controls the number of options to be displayed, or it may control the number of options displayed by intervening in the option display process without changing the custom AI. In the latter case, even if the custom AI outputs, for example, five options, the custom AI server 10 selects the number of options associated with the number of times the options have been displayed and displays them on the operator terminal 30. Note that because the "upper limit on the number of options displayed" is an upper limit, the custom AI server 10 may display a fewer number of options.

[0097] When options are displayed multiple times, they will gradually include options that have already been selected or options similar to them. By reducing the upper limit of options as described above, it is possible to make it less likely that such unnecessary options will be displayed compared to when the number of options displayed is fixed.

[0098] The custom AI server 10 may also increase the predetermined number in accordance with an increase in the number of analysis targets. In this case, the custom AI server 10 uses, for example, a number table that associates the number of analysis targets with the upper limit of the number of options to be displayed.

[0099] Figure 19 is a diagram showing another example of a number table. In the number table TB4 shown in Figure 19, when the "number of analysis targets" is "less than Th1," "greater than Th1 and less than Th2," or "greater than Th2," the "number of displayed options" is associated with "3 or less," "4," and "5 or more." The custom AI server 10 counts the number of types of user information that can be acquired using the API as the number of analysis targets, and displays the number of options associated with the count in the number table TB4.

[0100] In addition to "accounting information" and "financial information," analysis targets include, for example, "product information," "customer information," "personnel information," and "investor information." The granularity of the analysis target can be changed to include "balance sheets," "income statements," "cash flow statements," "statements of changes in shareholders' equity," and "annexed schedules" included in "accounting information." The more analysis targets there are, the more analysis methods that can be proposed. Therefore, by controlling the number of options displayed as described above, it becomes easier to suggest appropriate analysis even when the number of analysis targets increases, compared to when the number of displayed options is constant.

[0101] <Variation: Hierarchical option display> In the above example, an analysis method in which all options are displayed on the same level has been shown, but options may be displayed hierarchically. In this case, the option display operation may include an operation in which, when a first option and a second option are displayed together as options, if the first option is selected, the option corresponding to the first option is displayed as the next option, and if the second option is selected, the option corresponding to the second option is displayed as the next option.

[0102] The above-mentioned option display operation uses, for example, an option table in which options are defined for each level. FIG. 20 is a diagram showing an example of an option table. In option table TB5 shown in FIG. 20, options in the "first tier," options in the "second tier," and options in the "third tier" are shown in correspondence with one another. For example, "option A1" in the first tier is associated with multiple options in the second tier, such as "option A11," "option A12," and "option A13." Of these, "option A11" is associated with multiple options in the third tier, such as "option A111" and "option A112." Although FIG. 20 shows up to the third tier, the fourth and subsequent tiers may also be included.

[0103] For example, when "Option A1" (corresponding to the first option) in the first layer is selected, the custom AI displays "Option A11," "Option A12," "Option A13," etc. (options corresponding to the first option) in the second layer that are associated with "Option A1" as the next options. Furthermore, when "Option A12" (corresponding to the first option) is selected, the custom AI displays "Option A121," "Option A122," etc. (options corresponding to the first option) in the third layer that are associated with "Option A12" as the next options.

[0104] In addition, when "Option A2" (corresponding to the second option) in the first layer is selected, the custom AI displays "Option A21," "Option A22," "Option A23," etc. (options corresponding to the second option) in the second layer that correspond to "Option A2" as the next option.

[0105] The custom AI server 10 executes a reflection process to control the display of options based on the option table described above. Specifically, the custom AI server 10 executes a reflection process by including in the custom information a rule such as, for example, "When displaying options, refer to the option table TB5 and display options in order from the highest level. When an option is selected, select and display, as the next option, an option associated with the selected option from the options one level lower."

[0106] After performing a certain analysis, you may decide what you want to analyze next. For example, if you have performed an analysis on costs, you may want to analyze financial indicators that are highly correlated with costs. However, if there is only one level of options, the options in each level shown in Figure 20 are all in the same level, so the options may not be displayed in the same order. In the example above, you can control the order in which the options are selected, so the options can be displayed in a more orderly manner than if the options were not organized in a hierarchy.

[0107] <Examples of changes: businesses and operators> The business providing the custom AI server 10 and the business providing the service providing server 40 may be the same or different. If these business operators are different, the system operator using the operator terminal 30 may be the operator of the custom AI server 10 or the operator of the service providing server 40. In short, as long as the custom AI processing system 1 operates as described above, any system of business operators and operators may be used.

[0108] <Example of variation: Variation of composition> The configurations (overall configuration, hardware configuration, functional configuration, etc.) shown in FIG. 1 and elsewhere are merely examples, and other configurations may be used as long as they are not inconvenient for implementation. For example, the custom AI server 10 and the service providing server 40 may each be distributed across two or more devices (e.g., the AI ​​module 100 may be installed on another device), or may be provided in the form of SaaS (Software as a Service) or a cloud computing system. Furthermore, information processing performed by two or more information processing devices among the custom AI server 10, the user terminal 20, the operator terminal 30, and the service providing server 40 may be collectively executed by a device that integrates these devices. In short, as long as the necessary information processing is executed by the entire custom AI processing system 1, the devices that execute that information processing may have any configuration.

[0109] The output destination of information or data (hereinafter referred to as "information, etc.") may be another device, a display, a memory unit (including an internal memory unit and an external memory unit), an email address, an account of another system, etc. Acquisition of information, etc. includes acquiring information, etc. generated by the device itself, as well as acquiring information, etc. transmitted from another device. The table, etc. (table, database, etc.) in which parameters are associated is not limited to the illustrated table, etc., and the number of parameters may be reduced or increased. Furthermore, information, etc. corresponding to parameters may be obtained using a mathematical formula, a conditional formula, etc., without using a table, etc.

[0110] The above-described embodiments are information processing devices such as the custom AI server 10, the user terminal 20, and the operator terminal 30, and information processing systems such as the custom AI processing system 1 equipped with the custom AI server 10, the user terminal 20, and the operator terminal 30. However, they may also be information processing methods. Such information processing methods include the same steps as those executed by the information processing system. Furthermore, the above-described embodiments may also be programs. Such programs cause a computer to execute the same steps as those executed by the information processing system.

[0111] <Additional Notes> Furthermore, it may be provided in the following aspects.

[0112] (1) An information processing system having one or more processors, wherein the processor, in a reflection step, executes a reflection process for reflecting rule behavior in accordance with predetermined rules in an AI having a function of responding to instructions in natural language, wherein the rule behavior includes an input field display action for displaying an input field for instructions in natural language, an option generation action for generating option options for the instructions, an option display action for displaying the generated option, a response generation action for generating a response according to the instruction entered in the input field or the instruction indicated by the selected option, and a response display action for displaying the generated response, and the reflection process includes a provision process for providing rule information indicating the rule and samples of the option to the AI.

[0113] According to this embodiment, it is possible to easily give instructions to the AI.

[0114] (2) In the information processing system described in (1) above, the AI ​​has multiple types of analysis functions, the instructions are instructions that specify the information to be analyzed and the type of analysis function that will analyze the analysis target, the rule actions include an acquisition action that acquires the specified analysis target, and the response generation action includes an action that generates the results of analyzing the acquired analysis target with the specified type of analysis function as a response to the instructions.

[0115] According to this embodiment, it is possible to easily give instructions for analysis.

[0116] (3) In the information processing system described in (2) above, the reflection process includes a definition process that provides the AI ​​with definitions of terms used when specifying the analysis target and the analysis function.

[0117] According to this aspect, since inaccurate terms can cause erroneous analysis, the accuracy of the analysis can be improved by correctly defining terms.

[0118] (4) In the information processing system described in (2) or (3) above, the reflection process includes an explanation process that provides the AI ​​with an explanation of a method for analyzing the analysis target.

[0119] According to this aspect, it is possible to make it easier to obtain the intended analysis results.

[0120] (5) In the information processing system described in any one of (2) to (4) above, the providing process includes a process of providing the AI ​​with sample analysis instruction options including the added analysis target when information that can be obtained as the analysis target is added.

[0121] According to this aspect, the user can instruct analysis of the added analysis target without any effort.

[0122] (6) In the information processing system described in any one of (2) to (5) above, the providing process includes, when a new analysis function is added to the AI, providing the AI ​​with sample options for analysis instructions using the added function.

[0123] According to this aspect, the user can instruct analysis using the added analysis function without any effort.

[0124] (7) In the information processing system described in any one of (2) to (6) above, one or more parameters are used in the analysis instructions, the reflection process includes a candidate providing process that provides the AI ​​with candidate parameters, and the rule operation includes a parameter generation operation that generates a parameter to be used in the instructions indicated by the selected option, and a parameter display operation that displays the generated parameter.

[0125] According to this aspect, it is possible to easily specify parameters.

[0126] (8) In the information processing system according to any one of (2) to (7) above, the option display operation includes an operation of limiting the number of options to be displayed to a predetermined number or less.

[0127] According to this aspect, it is possible to prevent the situation where there are too many options and it becomes difficult to make a selection.

[0128] (9) In the information processing system described in (8) above, the processor decreases the predetermined number in accordance with an increase in the number of times the options are displayed.

[0129] According to this aspect, it is possible to make it difficult for unnecessary options to be displayed.

[0130] (10) In the information processing system described in (8) or (9) above, the processor increases the predetermined number in accordance with an increase in the number of analysis targets.

[0131] According to this aspect, it is possible to easily instruct appropriate analysis even if the number of analysis targets increases.

[0132] (11) In the information processing system described in any one of (1) to (10) above, the option display operation includes an operation of, when a first option and a second option are both displayed as the options, displaying an option corresponding to the first option as the next option when the first option is selected, and displaying an option corresponding to the second option as the next option when the second option is selected.

[0133] According to this aspect, the options can be displayed in order.

[0134] (12) An information processing method, in which a processor included in an information processing system executes each step of the information processing system described in any one of (1) to (10) above.

[0135] According to this embodiment, it is possible to easily give instructions to the AI.

[0136] (13) A program that causes a computer to execute each step of the information processing system according to any one of (1) to (10) above.

[0137] According to this embodiment, it is possible to easily give instructions to the AI. Of course, this is not the case. Furthermore, the above-described embodiments and modifications may be combined in any desired manner.

[0138] Finally, while various embodiments of the present invention have been described, these are presented by way of example only and are not intended to limit the scope of the invention. The novel embodiments may be embodied in various other forms, and various omissions, substitutions, and modifications may be made without departing from the spirit of the invention. The embodiments and their modifications are intended to be included within the scope and spirit of the invention, as well as within the scope of the inventions and their equivalents as defined in the appended claims. [Explanation of symbols]

[0139] 1: Custom AI processing system 2: Communication line 10: Custom AI Server 11: Control section 20: User terminal 21: Control unit 30: Operator terminal 31: Control unit 40: Service provider server 41: Control unit 100: AI module

Claims

1. An information processing system comprising one or more processors, the processor: In the reflection step, a reflection process is performed to reflect a rule operation according to a predetermined rule in the AI ​​having a function of responding to an instruction in natural language, The rule action includes: an input field display action that displays an input field for instructions in natural language; an option generating operation for generating an option for an instruction; an option display operation for displaying the generated options; a response generating operation of generating a response according to an instruction input into the input field or an instruction indicated by the selected option; and a response display operation for displaying the generated response; The reflection process includes a provision process of providing rule information indicating the rule and a sample of the option to the AI. Information processing system.

2. 2. The information processing system according to claim 1, The AI ​​has multiple types of analytical functions, the instruction is an instruction specifying information to be analyzed and a type of analysis function to analyze the information to be analyzed, The rule action includes an acquisition action for acquiring a designated analysis target; The response generation operation includes an operation of generating a result of analyzing the acquired analysis target using a specified type of analysis function as a response to the instruction. Information processing system.

3. 3. The information processing system according to claim 2, The reflection process includes a definition process of providing the AI ​​with definitions of terms used when specifying the analysis target and the analysis function. Information processing system.

4. 3. The information processing system according to claim 2, The reflection process includes an explanation process of providing the AI ​​with an explanation of how to analyze the analysis target. Information processing system.

5. 3. The information processing system according to claim 2, The providing process includes, when information that can be acquired as the analysis target is added, a process of providing the AI ​​with a sample of analysis instruction options including the added analysis target. Information processing system.

6. 3. The information processing system according to claim 2, The providing process includes, when a new analysis function is added to the AI, a process of providing the AI ​​with a sample of analysis instruction options using the added function. Information processing system.

7. 3. The information processing system according to claim 2, one or more parameters are used to direct the analysis; The reflection process includes a candidate providing process of providing the AI ​​with candidate parameters, The rule action includes a parameter generation action for generating a parameter to be used in an instruction indicated by the selected option, and a parameter display action for displaying the generated parameter. Information processing system.

8. 3. The information processing system according to claim 2, The option display operation includes an operation of limiting the number of options to be displayed to a predetermined number or less. Information processing system.

9. 9. The information processing system according to claim 8, the processor: The predetermined number is decreased in accordance with an increase in the number of times the options are displayed. Information processing system.

10. 9. The information processing system according to claim 8, the processor: Increasing the predetermined number in accordance with an increase in the number of analysis targets; Information processing system.

11. 2. The information processing system according to claim 1, The option display operation includes an operation of, when a first option and a second option are both displayed as the options, displaying an option corresponding to the first option as the next option when the first option is selected, and displaying an option corresponding to the second option as the next option when the second option is selected. Information processing system.

12. An information processing method, comprising: The processor of the information processing system Executing each step of the information processing system according to any one of claims 1 to 10. Information processing methods.

13. A program, A computer is caused to execute each step of the information processing system according to any one of claims 1 to 10. program.

Images