Information processing system, information processing method, and information processing program

Abstract

By prioritizing the use of information with high importance, it is possible to provide highly accurate responses to user requests, while by selecting and discarding information with low importance, it is possible to reduce processing load and increase response speed. [Solution] In the information processing system 10, the system terminal 12 is equipped with an information acquisition means capable of at least acquiring input information and user information input from an external terminal 16, an information classification means for determining whether specific information included in the input information or user information is of high importance to the user based on whether the specific information meets a predetermined standard, and if the specific information meets the predetermined standard, storing the information in a storage area as memory information, and if the information does not meet the predetermined standard, storing the information in a forgetting area as forgetting information, and an answer acquisition means for inputting the input information and memory information into a large-scale language model 15 and acquiring answer information to the input information and memory information from the large-scale language model.

Description

[Technical Field]

[0001] The present invention relates to techniques for generating answers to user prompts using large-scale language models (LLMs). [Background technology]

[0002] Conventionally, an information processing device has been proposed that is configured to improve the quality of answer data when answer data to question data is acquired using a large-scale language model (LLM) (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2025-65194 Summary of the Invention [Problem to be solved by the invention]

[0004] However, in recent years, there has been a demand for more accurate answers.

[0005] In view of the above situation, the present invention aims to provide an information processing system, an information processing method, and an information processing program that can provide highly accurate responses to user requests by prioritizing the use of information of high importance, and that can reduce processing load and increase response speed by selecting and discarding information of low importance. [Means for solving the problem]

[0006] The information processing system of the present invention is characterized by comprising: an information acquisition means capable of acquiring at least input information entered by a user and user information related to the user; an information classification means for determining whether specific information included in the input information or the user information is of high importance to the user based on whether the specific information meets a predetermined standard, and if the specific information meets the predetermined standard, storing the specific information in a storage area as memory information, and if the specific information does not meet the predetermined standard, storing the specific information in a forgetting area as forgotten information; and an answer acquisition means for inputting the input information and the memory information into a large-scale language model and acquiring answer information to the input information and the memory information from the large-scale language model.

[0007] Furthermore, the information processing method of the present invention is characterized by comprising: an information acquisition step capable of acquiring at least input information entered by a user and user information related to the user; an information classification step for determining whether specific information included in the input information or the user information is of high importance to the user based on whether the specific information meets a predetermined standard, and if the specific information meets the predetermined standard, storing the specific information in a storage area as memory information, and if the specific information does not meet the predetermined standard, storing the specific information in a forgetting area as forgotten information; and an answer acquisition step for inputting the input information and the memory information into a large-scale language model and acquiring answer information to the input information and the memory information from the large-scale language model.

[0008] Furthermore, an information processing program according to the present invention is an information processing program that causes a computer to function as: an information acquisition means that is capable of acquiring at least input information entered by a user and user information related to the user; an information classification means that determines whether specific information included in the input information or the user information is of high importance to the user based on whether the specific information meets a predetermined standard, and if the specific information meets the predetermined standard, stores the specific information in a storage area as memorized information, and if the specific information does not meet the predetermined standard, stores the specific information in a forgetting area as forgotten information; and an answer acquisition means that inputs the input information and the memorized information into a large-scale language model and acquires answer information to the input information and the memorized information from the large-scale language model. [Effects of the Invention]

[0009] The information processing system, information processing method, and information processing program according to the present invention can provide highly accurate responses to user requests by prioritizing the use of information of high importance, and can also achieve the excellent effect of reducing processing load and increasing response speed by selecting and discarding information of low importance. [Brief explanation of the drawings]

[0010] [Figure 1] 1 is a schematic diagram showing an overview of an information processing system 10 according to the present embodiment. [Figure 2] 1 is a system configuration diagram showing an example of the configuration of an information processing system 10 according to the present embodiment. [Figure 3] 1A and 1B are diagrams showing an example of a memory log MLG and an example of a forgetting log ELG, respectively, when the frequency of use of specific information in a predetermined period is used as an index for judgment. [Figure 4]1A and 1B are diagrams showing an example of a memory log MLG and an example of a forgetting log ELG, respectively, in which the elapsed time since a specific piece of information was last used is used as an index for judgment. [Figure 5] (a-1) A diagram showing an example of a memory log MLG when the frequency of use of specific information in multiple sentences is used as an index for judgment. (a-2) A diagram showing an example of a forgetting log ELG when the frequency of use of specific information in multiple sentences is used as an index for judgment. (b-1) A diagram showing an example of a memory log MLG when the association between specific information and user information USI is used as an index for judgment. (b-2) A diagram showing an example of a forgetting log ELG when the association between specific information and user information USI is used as an index for judgment. DETAILED DESCRIPTION OF THE INVENTION

[0011] An information processing system 10 according to an embodiment of the present invention will be described below with reference to the drawings.

[0012] <System Overview> First, an overview of an information processing system 10 according to this embodiment will be described with reference to Fig. 1. Fig. 1 is a schematic diagram showing an overview of the information processing system 10 according to this embodiment.

[0013] The information processing system 10 is a system comprising: an information acquisition means 12a capable of acquiring at least input information INI entered by a user and user information USI related to the user; an information classification means 12b that determines whether specific information included in the input information INI or the user information USI is of high importance to the user based on whether the specific information meets a predetermined standard; and, if the specific information meets the predetermined standard, stores the specific information as memorized information MEI in a memory log MLG; and, if the specific information does not meet the predetermined standard, stores the specific information as forgotten information ERI in a forgetting log ELG; and an answer acquisition means 12c that inputs the input information INI and memorized information MEI into a large-scale language model (LLM) and acquires answer information ASI for the input information INI and the memorized information MEI from the large-scale language model.

[0014] According to the information processing system 10, by prioritizing the use of information of high importance, it is possible to provide highly accurate responses to user requests, and by selecting and discarding information of low importance, it is possible to reduce processing load and increase response speed.

[0015] Furthermore, regardless of whether the specific information contained in the input information INI or the user information USI meets a predetermined standard, if there is a command in the input information INI suggesting that the specific information be stored in the memory log MLG as memory information MEI, the information classification means 12b stores the specific information in the memory log MLG, and if there is a command in the input information INI suggesting that the specific information be deleted from the forgetting log ELG, the information classification means 12b deletes the specific information from the forgetting log ELG.

[0016] With this configuration, the user himself / herself can classify remembered information and forgotten information, which can further improve the accuracy of the answers and increase convenience for the user.

[0017] In addition, the system is configured so that each user can view the memory log MLG and the forgetting log ELG.

[0018] With this configuration, the user can check the contents of the memorized information and forgotten information and then select which information to keep and which to forget, thereby further improving the accuracy of the answers and increasing convenience for the user.

[0019] Furthermore, the information classification means 12b periodically deletes stored information MEI that does not meet a certain standard from the storage log MLG.

[0020] With this configuration, by deleting information of low importance, it is possible to reduce the processing load and increase the response speed.

[0021] <System configuration example> Next, a configuration example of the information processing system 10 according to this embodiment will be described with reference to Fig. 2. Fig. 2 is a system configuration diagram showing a configuration example of the information processing system 10 according to this embodiment.

[0022] The information processing system 10 can be configured, for example, to include a system terminal 12 that controls the entire system, an external storage 14, an LLM (large scale language model) 15, and an external terminal 16 that are connected to the system terminal 12 via a network NW so that they can communicate with each other.

[0023] The system terminal 12 is a terminal that controls the entire information processing system 10, and is configured from a conventionally known server, personal computer, etc. In this example, the system terminal 12 is configured from one server, but it may be configured from multiple servers, personal computers, etc. The hardware configuration of the system terminal 12 and the programs executed by the system terminal 12 will be described later.

[0024] The external storage 14 is a storage means capable of saving (storing) the user information USI, the memory log MLG, and the forgetting log ELG, and in this example is configured as a server equipped with a large-capacity storage medium. Note that instead of providing the external storage 14, the storage device 26 provided in the system terminal 12 may be used as internal storage for saving (storing) the user information USI, the memory log MLG, and the forgetting log ELG.

[0025] The LLM 15 is a large-scale language model of multiple types, and in this example, is configured by a server that configures the large-scale language model. Note that the LLM 14 may be a local LLM or a public LLM (cloud LLM).

[0026] The external terminal 16 is a terminal used by a user (an individual or a specific group consisting of multiple people (a team, a department, a company, etc.)) who uses the information processing system 10, and is composed of a personal computer, a tablet, a smartphone, etc. The type of the external terminal 16 is not particularly limited, but includes, for example, a smartphone, a personal computer, a tablet used by an individual, or a smartphone, a personal computer, a tablet, etc. used by a company employee, etc.

[0027] The network NW is a line that allows the system terminal 12, external storage 14, LLM 15, and external terminal 16 to communicate with each other, and is typically configured as a WAN (Wide Area Network), also known as the Internet. The network NW may be wired or wireless, may be a LAN (Local Area Network), may be a dedicated line such as a VPN (Virtual Private Network), or may be a combination of these lines.

[0028] <System terminal / hardware configuration example> Next, an example of the hardware configuration of the system terminal 12 will be described.

[0029] As shown in FIG. 2, the system terminal 12 is configured to include, for example, a CPU 21, a ROM 22, a RAM 23, an external storage drive 25, a storage device 26, an input device 27, a display device 28, a communication unit 29, etc., all of which are connected to the CPU 21 via a bus.

[0030] The CPU 21 is a control means for controlling the entire system terminal 12, and performs processes such as executing application programs and operating systems (OS) stored in the ROM 22, storage device 26, etc., and storing data and files necessary for executing programs in the RAM 23, storage device 26, etc.

[0031] The ROM 22 is a storage means for storing basic I / O programs and various data, and is configured, for example, by a PROM, flash memory, etc. The RAM 23 is a storage means for temporarily storing data, and is configured, for example, by an SDRAM, DRAM, etc. The external storage drive 25 is a control means capable of reading and writing data from and to a recording medium 24 such as a magnetic tape or DVD, and is configured, for example, by a magnetic tape storage, DVD drive, etc.

[0032] The storage device 26 is a storage means for storing application programs, OS, control programs, related programs, various information, etc., and is configured, for example, by a hard disk (HDD), solid state drive (SDD), etc. The input device 27 is used to input commands (instructions) etc. to the system terminal 12, and is configured, for example, by a keyboard, a pointing device (mouse, etc.), a touch panel, etc.

[0033] The display device 28 displays commands input by the input device 27, response outputs from the system terminal 12 to those commands, various displays, etc., and is configured, for example, by a liquid crystal display device, a plasma display, an organic EL display, etc. The communication unit 29 is control means that controls communication with the external storage 14, the LLM 15, the external terminal 16, etc. via the network NW, and is configured, for example, by a communication card, etc.

[0034] <Management terminal / function> Next, the functions of the system terminal 12 will be described.

[0035] The storage device 26 of the system terminal 12 stores a program (information processing program) that causes the system terminal 12 to function as an information acquisition means 12a, an information classification means 12b, and an answer acquisition means 12c.

[0036] <Management terminal / function / information acquisition method> Next, the information acquisition means 12a will be described.

[0037] The information acquisition means 12a is a means capable of acquiring at least input information INI entered by a user (an individual or a specific group consisting of multiple people (a team, a department, a company, etc.)) and user information USI related to the user, and in this example, is composed of a program stored in the memory device 26 of the system terminal 12, the memory device 26, etc.

[0038] The information acquiring means 12a acquires information input by a user of the information processing system 10 using the external terminal 16 as input information INI, and stores the acquired input information INI in the storage device 26 in association with the user's ID.

[0039] Here, "input information INI" refers to information (data) input by a user of the information processing system 10 using the input means (keyboard, pointing device (mouse, etc.), touch panel, microphone, etc.) of the external terminal 16. The data format of the "input information INI" is not particularly limited, and may be, for example, text data, audio data, still image data, video data, files, etc. Furthermore, the input information INI is not limited to information input by an individual, but may also be information input by a specific group consisting of multiple people (a team, department, company, etc.) (for example, information obtained through an in-company chat or document sharing system, etc.).

[0040] Examples of "input information INI" include text data and voice data such as a user's ID and password, questions and prompts for the information processing system 10, files uploaded by the user, still image data, or video data. When registering an account for the information processing system 10, the information processing system 10 has the user input user information USI using the external terminal 16, and stores the input user information USI in the external storage 14 in association with the user's ID. Furthermore, when using the information processing system 10, the information processing system 10 appropriately acquires user information USI included in input information IN, and stores the acquired user information USI in the external storage 14 in association with the user's ID.

[0041] Here, "user information USI" refers to information relating to a user of the information processing system 10, and in this example, is stored in the external storage 14 in association with the user's ID. It may also be stored in the storage device 26 in association with the user's ID.

[0042] "User information USI" includes, for example, the user's personal information (e.g., age, gender, place of employment, job title, qualifications held, field of expertise, purpose of using the system, etc.), the user's genre preferences (e.g., medicine, management, law, history, economics, politics, etc.), the user's knowledge level (e.g., use of terminology and question format), and the user's vocabulary style (e.g., extensive use of technical terms, abstract questions, specific instructions, etc.).

[0043] After acquiring the input information INI, the information acquisition means 12a acquires user information USI associated with the user ID included in the input information INI from the user information USI stored in the external storage 14, and stores the acquired user information USI in the storage device 26 in association with the user ID.

[0044] <Management terminal / function / information classification method> Next, the information classification means 12b will be described.

[0045] The information classification means 12b is a means for determining whether specific information contained in the input information INI or user information USI acquired by the information acquisition means 12a is of high importance to the user based on whether the specific information meets a predetermined standard, and if the specific information meets the predetermined standard, stores the specific information in the memory log MLG as memory information MEI, and if the specific information does not meet the predetermined standard, stores the specific information in the forgetting log ELG as forgetting information ERI, and in this example, is composed of a program stored in the memory device 26 of the system terminal 12, the memory device 26, etc.

[0046] First, the information classification means 12b acquires the input information INI and user information USI stored in the storage device 26 by the information acquisition means 12a, and determines whether the information contained in this input information INI or user information USI meets a predetermined standard.

[0047] Here, indicators for "determining whether a specified standard is met" include, for example, the frequency of use of specific information in a specified period, the time elapsed since the specific information was last used, the frequency of use of specific information in multiple sentences, and the relevance of specific information to user information USI. Note that the "determination of whether a specified standard is reached" is not limited to a binary determination of whether the indicator is above or below a specified threshold. For example, multiple levels of standards may be set, such as a first standard and a second standard, where (condition 1) the indicator is above the first standard, (condition 2) the indicator is below the first standard but above or below the second standard, or (condition 3) the indicator is below the second standard, and if (condition 1) is met, it may be determined that the specified standard is reached, and if (condition 3) is met, it may be determined that the specified standard is not reached. Furthermore, the indicators for "determining whether a specified standard is met" may be indicators relating to an individual's attributes, behavior, preferences, etc., or may be indicators relating to the attributes, behavior, preferences, etc. of a specific group consisting of multiple people (a team, department, company, etc.) (for example, attribute information, behavioral history, etc. of the entire group).

[0048] Next, if specific information included in the input information INI or the user information USI meets a predetermined standard, the information classification means 12b determines that the information is of high importance to the user, assigns a high score to the specific information, and stores the specific information as stored information MEI in a storage log MLG provided for each user. Note that the storage log MLG may be any storage area (storage area) capable of storing the stored information MEI, and the storage format is not limited to a log. Furthermore, the stored information MEI is not limited to information stored by an individual, but may also be information stored by a specific group consisting of multiple people (a team, a department, a company, etc.).

[0049] On the other hand, if specific information included in the input information INI or the user information USI does not meet a predetermined standard, it is determined that the information is of low importance to the user, a low score is assigned to the specific information, and the specific information is stored as forgotten information ERI in a forgetting log ELG provided for each user. Note that the forgetting log ELG may be any storage area (forgetting area) capable of storing forgotten information ERI, and the storage format is not limited to a log. Furthermore, forgotten information ERI is not limited to information stored by an individual, but may also be information stored by a specific group consisting of multiple people (a team, department, company, etc.).

[0050] For example, when "frequency of use of specific information in a specified period" is used as an index for judgment, the specific information contained in the input information INI is extracted, and the number of times the specific information was used in the specified period is counted. If the number of times the specific information was used is greater than a specified standard, the information is determined to be of high importance to the user and a high score is assigned. The specific information is then stored in the memory log MLG as memory information MEI, associated with the user's ID.

[0051] On the other hand, if the number of times the information is used is less than a predetermined standard, the information is determined to be of low importance to the user and a low score is assigned. The specific information is then stored in the forgetting log ELG as forgotten information ERI, associated with the user's ID, and so on.

[0052] Figure 3(a) is a diagram showing an example of a memory log MLG when the frequency of use of specific information in a specified period is used as an index for judgment, and Figure 3(b) is a diagram showing an example of a forgetting log ELG when the frequency of use of specific information in a specified period is used as an index for judgment.

[0053] As shown in Figure 3(a), if the number of times specific information X1 (e.g., a specific keyword) is used in a predetermined period T1 (e.g., 10 days) is more than a predetermined standard (e.g., 4 times) (e.g., 10 times), the information is determined to be of high importance to the user, and the score of the frequency of use of the specific information X1 in the predetermined period T1 is assigned, for example, 10 points out of 10, and the specific information X1 is stored as memory information MEI in the memory log MLG in association with the user's ID.

[0054] Furthermore, if the number of times that specific information X2 (e.g., a specific keyword) is used in a predetermined period T1 (e.g., 10 days) is greater than a predetermined standard (e.g., 4 times) (e.g., 5 times), the information is determined to be of high importance to the user, and the score for the frequency of use of the specific information X2 in the predetermined period T1 is assigned, for example, 5 points out of 10, and the specific information X2 is stored as memory information MEI in the memory log MLG in association with the user's ID.

[0055] On the other hand, as shown in Figure 3(b), if the number of times that specific information X3 (e.g., a specific keyword) is used in a predetermined period T1 (e.g., 10 days) is less than a predetermined standard (e.g., 4 times) (e.g., 1 time), the information is determined to be of low importance to the user, and the score for the frequency of use of the specific information X3 in the predetermined period T1 is assigned, for example, 1 point out of 10, and the specific information is stored in the forgetting log ELG as forgotten information ERI, associated with the user's ID.

[0056] When "the time elapsed since the specific information was last used" is used as an indicator for judgment, the specific information contained in the input information INI is extracted, and if the time elapsed since the specific information was last used is shorter than a predetermined standard, it is determined that the information is of high importance to the user, a high score is assigned to the specific information, and the specific information is stored as memory information MEI in the memory log MLG, associated with the user's ID.

[0057] On the other hand, if the time elapsed since the specific information was last used is longer than a predetermined standard, the information is determined to be of low importance to the user, a low score is assigned to the specific information, and the specific information is stored in the forgetting log ELG as forgotten information ERI, associated with the user's ID.

[0058] Figure 4(a) is a diagram showing an example of a memory log MLG when the elapsed time since specific information was last used is used as an indicator for judgment, and Figure 4(b) is a diagram showing an example of a forgetting log ELG when the elapsed time since specific information was last used is used as an indicator for judgment.

[0059] As shown in Figure 4(a), if the time that has elapsed since specific information Y1 (for example, a specific keyword) was last used is shorter than a predetermined standard (for example, if the specific information Y1 was last used one day ago, which is shorter than five days ago), the information is determined to be of high importance to the user, and a score of, for example, 8 points out of 10 is assigned to the time that has elapsed since the specific information Y1 was last used, and the specific information Y1 is stored as memory information MEI in the memory log MLG, associated with the user's ID.

[0060] Furthermore, if the time that has elapsed since specific information Y2 (for example, a specific keyword) was last used is shorter than a predetermined standard (for example, if the specific information Y2 was last used four days ago, which is shorter than five days ago), the information is determined to be of high importance to the user, and a score of, for example, 3 points out of 10 is assigned to the time that has elapsed since the specific information Y2 was last used, and the specific information Y2 is stored as memory information MEI in the memory log MLG, associated with the user's ID.

[0061] On the other hand, as shown in Figure 4(b), if the time that has elapsed since specific information Y3 (e.g., a specific keyword) was last used is longer than a predetermined standard (e.g., the most recent time that specific information Y3 was used was 7 days ago, which is longer than 5 days ago), a score of 1 point out of 10 is assigned to the time that has elapsed since the specific information Y3 was last used, and the specific information Y3 is stored in the forgetting log ELG as forgotten information ERI, associated with the user's ID.

[0062] When "frequency of use of specific information in multiple sentences" is used as an index for judgment, the specific information contained in the input information INI is extracted, and the number of times the specific information is used in multiple sentences is counted. If the number of times the specific information is used is greater than a predetermined standard, the information is judged to be of high importance to the user, a high score is assigned, and the specific information is stored in the memory log MLG as memory information MEI, associated with the user's ID.

[0063] On the other hand, if the number of times the information is used is less than a predetermined standard, the information is determined to be of low importance to the user, a low score is assigned, and the specific information is stored in the forgetting log ELG as forgotten information ERI, associated with the user's ID.

[0064] Figure 5(a-1) is a diagram showing an example of a memory log MLG when the frequency of use of specific information in multiple sentences is used as an index for judgment, and Figure 5(a-2) is a diagram showing an example of a forgetting log ELG when the frequency of use of specific information in multiple sentences is used as an index for judgment.

[0065] As shown in Figure 5(a-1), if the number of times that specific information Z (e.g., a specific keyword) is used in three consecutive sentences 1 to 3 is more (e.g., 5 times) than a predetermined standard (e.g., 3 times), it is determined that the information is of high importance to the user, and the score for the frequency of use of the specific information Z in multiple sentences is assigned, for example, 9 points out of 10, and the specific information Z is stored as memory information MEI in the memory log MLG in association with the user's ID.

[0066] On the other hand, as shown in Figure 5(a-2), if the number of times specific information Z is used in three consecutive sentences 1 to 3 is less than a predetermined standard (e.g., 3 times) (e.g., 0 times), the information is determined to be of low importance to the user, and the score for the frequency of use of the specific information Z in multiple sentences is assigned, for example, 0 points out of 10, and the specific information Z is stored in the forgetting log ELG as forgotten information ERI, associated with the user's ID.

[0067] When the "relationship between specific information and user information USI" is used as an indicator for judgment, specific information that is identical or similar to the user information USI is extracted from the input information INI, the number of pieces of specific information is counted, and if the number is greater than a predetermined standard, it is determined that the information is of high importance to the user, a high score is assigned to the specific information, and the specific information is stored as memory information MEI in the memory log MLG, associated with the user's ID.

[0068] On the other hand, if the number is less than a predetermined standard, the information is determined to be of low importance to the user, a low score is assigned to the specific information, and the specific information is stored in the forgetting log ELG as forgotten information ERI, associated with the user's ID.

[0069] Figure 5(b-1) is a diagram showing an example of a memory log MLG when the association between specific information and user information USI is used as an index for judgment, and Figure 5(b-2) is a diagram showing an example of a forgetting log ELG when the association between specific information and user information USI is used as an index for judgment.

[0070] Figure 5(a-1) is a diagram showing an example of a memory log MLG when the frequency of use of specific information in multiple sentences is used as an index for judgment, and Figure 5(a-2) is a diagram showing an example of a forgetting log ELG when the frequency of use of specific information in multiple sentences is used as an index for judgment.

[0071] For example, as shown in Figure 5(b-1), if the number of specific information identical to the user information USI (e.g., the word "economy") is greater than a predetermined standard (e.g., three) (e.g., four), the information is determined to be of high importance to the user, and a score of 10 out of 10 is assigned as the score of the relevance between the specific information and the user information USI, and the specific information (the word "economy") is stored as memory information MEI in the memory log MLG in association with the user's ID.

[0072] On the other hand, if the number of specific pieces of information identical to the user information USI is less than a predetermined standard (e.g., 3 pieces) (e.g., 0 pieces), the information is determined to be of low importance to the user, and a score of 0 out of 10 is assigned as the score of the relevance between the specific information and the user information, and the specific information (the word "economy") is stored in the forgetting log ELG as forgotten information ERI, associated with the user's ID.

[0073] Furthermore, regardless of whether specific information contained in the input information INI or the user information USI meets a predetermined standard, if there is a command in the input information INI suggesting that specific information be stored in the memory log MLG as memory information MEI, the information classification means 12b stores the specific information in the memory log MLG, and if there is a command in the input information INI suggesting that the memory information be deleted from the memory log MLG, the information classification means 12b deletes the memory information from the memory log MLG.

[0074] Here, "a command suggesting that specific information in the input information INI be stored in the memory log MLG as memory information MEI" corresponds to, for example, if the "specific information" is the string XYZ, inputting text data such as "remember XYZ," "memorize XYZ," or "record XYZ" into a prompt, which is one of the pieces of input information INI, using the external terminal 16.

[0075] Furthermore, "a command in the input information INI suggesting that the stored information MEI be deleted from the stored log MLG" corresponds to, for example, when the "stored information MEI" is the string ABC, inputting text data such as "Forget ABC," "Delete ABC," or "Erase ABC" into a prompt, which is one of the pieces of input information INI, using the external terminal 16.

[0076] For example, in the example described using Figure 3(a), specific keywords X1 and X2 are stored in the memory log MLG as memory information MEI, but when text data such as "remember X3," "memorize X3," or "record X3" is entered in a prompt, which is one of the input information INI, using the external terminal 16, the information classification means 12b adds and stores the specific keyword X3 in the memory log MLG.

[0077] On the other hand, when text data such as "Forget X2," "Delete X2," or "Erase X2" is entered in the prompt, which is one of the input information INI, using the external terminal 16, the specific keyword X2 is deleted from the memory log MLG.

[0078] According to this example, the user himself / herself can classify remembered information and forgotten information, which can further improve the accuracy of the answer and increase the convenience for the user.

[0079] In this example, the memory log MLG and the forgetting log ELG are configured to be viewable for each user.

[0080] For example, in the example described using Figure 3(a), specific keywords X1 and X2 are stored in the memory log MLG as memory information MEI, and when there is a request from the external terminal 16 to view the memory log MLG, the information processing system 10 transmits information on the specific keywords X1 and X2 stored in the memory log MLG to the external terminal 16, making it possible to display it on the external terminal 16.

[0081] In addition, in the example described using Figure 3(b), a specific keyword X3 is stored in the forgetting log ELG as forgetting information ERI, and when there is a request from the external terminal 16 to view the forgetting log ELG, the information processing system 10 transmits information on the specific keyword X3 stored in the forgetting log ELG to the external terminal 16, making it possible to display it on the external terminal 16.

[0082] According to this example, after checking the contents of the memorized information and forgotten information, the user can select which information to keep and which to forget, thereby further improving the accuracy of the answers and increasing convenience for the user.

[0083] The information classification means 12b is also configured to periodically delete stored information MEI that does not meet a certain standard from the storage log MLG.

[0084] For example, in the example described using Figure 3(a), the memory log MLG stores a specific keyword X1 (score 10) and a specific keyword X2 (score 5) as memory information MEI, but the information classification means 12b screens the memory log MLG at a predetermined period and deletes memory information MEI (in this example, the specific keyword X2) that does not meet a certain standard (for example, score 6) from the memory log MLG.

[0085] Also, in the example described using Figure 4(a), the memory log MLG stores a specific keyword Y1 (score 8) and a specific keyword X2 (score 3) as memory information MEI, but the information classification means 12b screens the memory log MLG at a predetermined period and deletes memory information MEI (in this example, the specific keyword Y2) that does not meet a certain standard (for example, score 4) from the memory log MLG.

[0086] According to this example, by periodically deleting information of low importance, it is possible to reduce the processing load and increase the response speed.

[0087] <Management device / function / answer acquisition method> Next, the answer acquisition means 12c will be described.

[0088] The answer acquisition means 12c is a means for inputting input information INI and stored information MEI into a large-scale language model (LLM) and acquiring answer information ASI for the input information INI and stored information MEI from the large-scale language model, and in this example is composed of a program stored in the memory device 26 of the system terminal 12, the memory device 26, etc.

[0089] The answer acquisition means 12c inputs the input information INI acquired by the information acquisition means 12a and the stored information MEI stored in the storage device 26 by the information classification means 12b into a large-scale language model (LLM) LLM15, acquires answer information for the input information INI and the stored information MEI from the LLM15, and then transmits the answer information ASI to the external terminal 16.

[0090] For example, the answer acquisition means 12c inputs (transmits) input information INI (e.g., a command to translate English into Japanese and English text data) and stored information MEI (e.g., frequently used Japanese words) to the LLM 15, and transmits the answer generated by the LLM 15 (e.g., a Japanese sentence that prioritizes frequently used Japanese words) to the external terminal 16 as answer information ASI.

[0091] In addition, the answer acquisition means 12c inputs (transmits) input information INI (e.g., a command to request the latest news from around the world) and stored information MEI (e.g., the name of a specific country) to the LLM 15, and transmits the answer generated by the LLM 15 (e.g., the latest news from around the world searched with a specific country as the priority) to the external terminal 16 as answer information ASI.

[0092] In addition, the answer acquisition means 12c inputs (transmits) input information INI (e.g., a command to request the latest news in Japan) and stored information MEI (e.g., a user's personal information) to the LLM 15, and transmits the answer generated by the LLM 15 (e.g., the latest news in Japan that matches the user's personal information) to the external terminal 16 as answer information ASI.

[0093] In addition, the answer acquisition means 12c inputs (transmits) input information INI (e.g., a command to generate a specified image) and stored information MEI (e.g., a knowledge level) to the LLM 15, and transmits the answer generated by the LLM 15 (e.g., image data of an image that matches the user's knowledge level) to the external terminal 16 as answer information ASI.

[0094] <Information Processing System / Summary> As described above, the information processing system according to this embodiment (for example, the information processing system 10 shown in FIGS. 1 and 2) includes an information acquisition unit (for example, the information acquisition unit 12a shown in FIGS. 1 and 2) that can acquire at least input information input by a user (for example, the input information INI shown in FIGS. 1 and 2) and user information related to the user (for example, the user information USI shown in FIGS. 1 and 2), and a function that determines whether specific information included in the input information or the user information is highly important to the user based on whether the specific information meets a predetermined standard, and if the specific information meets the predetermined standard, stores the specific information as stored information (for example, the stored information MEI shown in FIGS. 1 and 2). and if the specific information does not reach the predetermined standard, stores the specific information in a forgetting area (for example, the forgetting log ELG shown in FIGS. 1 and 2) as forgotten information (for example, the forgotten information ERI shown in FIGS. 1 and 2); and answer acquisition means (for example, the answer acquisition means 12c shown in FIGS. 1 and 2) that inputs the input information and the stored information into a large-scale language model and acquires answer information (for example, the answer information ASI shown in FIGS. 1 and 2) to the input information and the stored information from the large-scale language model.

[0095] According to the information processing system of this embodiment, by prioritizing the use of information of high importance, it is possible to provide highly accurate responses to user requests, and by selecting and discarding information of low importance, it is possible to reduce processing load and increase response speed.

[0096] In addition, the information classification means may store the specific information in the storage area if there is an instruction in the input information suggesting that the specific information contained in the input information or the user information should be stored in the storage area as the storage information, regardless of whether the specific information meets the specified standard, and may delete the storage information from the storage area if there is an instruction in the input information suggesting that the storage information should be deleted from the storage area.

[0097] With this configuration, the user himself / herself can classify remembered information and forgotten information, which can further improve the accuracy of the answers and increase convenience for the user.

[0098] The storage area and the forgotten area may be configured to be viewable for each user.

[0099] With this configuration, the user can check the contents of the memorized information and forgotten information and then select which information to keep and which to forget, thereby further improving the accuracy of the answers and increasing convenience for the user.

[0100] The information classification means may periodically delete the stored information that does not meet a certain standard from the storage area.

[0101] With this configuration, by deleting information of low importance, it is possible to reduce the processing load and increase the response speed.

[0102] Furthermore, the information processing method according to this embodiment (for example, a method executed by the information processing system 10 shown in FIGS. 1 and 2) includes an information acquisition step (for example, processing executed by the information acquisition means 12a shown in FIGS. 1 and 2) capable of acquiring at least input information input by a user (for example, input information INI shown in FIGS. 1 and 2) and user information related to the user (for example, user information USI shown in FIGS. 1 and 2), and a step of determining whether specific information included in the input information or the user information is highly important to the user based on whether the specific information meets a predetermined standard, and if the specific information meets the predetermined standard, storing the specific information as stored information (for example, stored information MEI shown in FIGS. 1 and 2). and if the specific information does not meet the predetermined standard, stores the specific information in a forgetting area (for example, the forgetting log ELG shown in FIGS. 1 and 2) as forgotten information (for example, the forgotten information ERI shown in FIGS. 1 and 2); and an answer acquisition step (for example, a process executed by the answer acquisition means 12c shown in FIGS. 1 and 2) of inputting the input information and the stored information into a large-scale language model and acquiring answer information (for example, the answer information ASI shown in FIGS. 1 and 2) to the input information and the stored information from the large-scale language model.

[0103] According to the information processing method of this embodiment, by prioritizing the use of information of high importance, it is possible to provide highly accurate responses to user requests, and by selecting and discarding information of low importance, it is possible to reduce processing load and increase response speed.

[0104] Furthermore, an information processing program according to this embodiment (for example, a program executed by the information processing system 10 shown in FIGS. 1 and 2) includes a computer, an information acquisition unit (for example, information acquisition unit 12a shown in FIGS. 1 and 2) that can acquire at least input information input by a user (for example, input information INI shown in FIGS. 1 and 2) and user information related to the user (for example, user information USI shown in FIGS. 1 and 2), and a function that determines whether specific information included in the input information or the user information is highly important to the user based on whether the specific information meets a predetermined standard, and if the specific information meets the predetermined standard, stores the specific information in stored information (for example, stored information USI shown in FIGS. 1 and 2). and if the specific information does not reach the predetermined standard, stores the specific information in a forgetting area (for example, the forgetting log ELG shown in FIGS. 1 and 2) as forgotten information (for example, forgotten information ERI shown in FIGS. 1 and 2); and functioning as answer acquisition means (for example, the answer acquisition means 12c shown in FIGS. 1 and 2) that inputs the input information and the stored information into a large-scale language model and acquires answer information (for example, answer information ASI shown in FIGS. 1 and 2) to the input information and the stored information from the large-scale language model.

[0105] According to the information processing program of this embodiment, by prioritizing the use of information of high importance, it is possible to provide highly accurate responses to user requests, and by selecting and discarding information of low importance, it is possible to reduce processing load and increase response speed.

[0106] It should be noted that the information processing system, information processing method, and information processing program according to the present invention are not limited to the above-described embodiments, and various modifications can of course be made within the scope that does not deviate from the gist of the present invention.

[0107] Therefore, for example, in the above example, the information classification means stores the specific information contained in the input information or the user information in the storage area as the memory information when there is a command in the input information suggesting that the specific information be stored in the storage area, regardless of whether the specific information meets the specified standard, and deletes the memory information from the storage area when there is a command in the input information suggesting that the memory information be deleted from the storage area.However, the information classification means may also store the specific information in the forgetting area when there is a command in the input information suggesting that the specific information be stored in the forgetting area as the forgetting information, regardless of whether the specific information contained in the input information or the user information meets the specified standard, and delete the forgetting information from the forgetting area when there is a command in the input information suggesting that the forgetting information be deleted from the forgetting area. [Industrial Applicability]

[0108] The information processing system, information processing method, and information processing program according to the present invention can be widely applied to various fields as an interface for local LLM or cloud LLM (open LLM). [Explanation of symbols]

[0109] 10 Information Processing Systems 12 System Terminal 12a Information acquisition means 12b Information classification means 12c Means of obtaining answers 14 External Storage 16 External Terminal 21 CPU 22 ROM 23 RAM 24 Recording media 25 External Storage Drive 26 Storage device 27 Input Devices 28 Display device 29 Communications Department NW Network INI input information USI User Information MEI Memory Information MLG Memory Log ERI forgotten information ELG Forgotten Log ASI response information

Claims

1. an information acquisition means capable of acquiring at least input information input by a user and user information relating to the user; an information classification means for determining whether specific information included in the input information or the user information is highly important to the user based on whether the specific information meets a predetermined standard, and if the specific information meets the predetermined standard, storing the specific information in a storage area as storage information, and if the specific information does not meet the predetermined standard, storing the specific information in a forgetting area as forgetting information; and an answer acquisition means for inputting the input information and the stored information into a large-scale language model and acquiring answer information for the input information and the stored information from the large-scale language model. An information processing system comprising:

2. 2. The information processing system according to claim 1, the information classification means stores the specific information in the storage area when there is a command in the input information suggesting that the specific information included in the input information or the user information should be stored as the storage information, regardless of whether the specific information meets the predetermined standard, and deletes the storage information from the storage area when there is a command in the input information suggesting that the storage information should be deleted from the storage area. An information processing system comprising:

3. 3. The information processing system according to claim 2, The storage area and the forgotten area are configured to be viewable for each user. An information processing system comprising:

4. 3. The information processing system according to claim 1, the information classification means periodically deletes the stored information that does not meet a certain standard from the storage area; An information processing system comprising:

5. an information acquisition step capable of acquiring at least input information input by a user and user information related to the user; an information classification step of determining whether specific information included in the input information or the user information is highly important to the user based on whether the specific information meets a predetermined standard, and if the specific information meets the predetermined standard, storing the specific information in a storage area as storage information, and if the specific information does not meet the predetermined standard, storing the specific information in a forgetting area as forgetting information; an answer acquisition step of inputting the input information and the stored information into a large-scale language model and acquiring answer information to the input information and the stored information from the large-scale language model; 1. An information processing method comprising:

6. 6. The information processing method according to claim 5, The information classification step stores the specific information in the storage area when there is an instruction in the input information suggesting that the specific information included in the input information or the user information should be stored as the storage information, regardless of whether the specific information meets the predetermined standard, and deletes the storage information from the storage area when there is an instruction in the input information suggesting that the storage information should be deleted from the storage area.

1. An information processing method comprising:

7. Computer, an information acquisition means capable of acquiring at least input information input by a user and user information relating to the user; an information classification means for determining whether specific information included in the input information or the user information is highly important to the user based on whether the specific information meets a predetermined standard, and if the specific information meets the predetermined standard, storing the specific information in a storage area as storage information, and if the specific information does not meet the predetermined standard, storing the specific information in a forgetting area as forgetting information; inputting the input information and the stored information into a large-scale language model, and functioning as an answer acquisition means for acquiring answer information to the input information and the stored information from the large-scale language model; An information processing program characterized by:

8. 8. The information processing program according to claim 7, the information classification means stores the specific information in the storage area when there is a command in the input information suggesting that the specific information included in the input information or the user information should be stored as the storage information, regardless of whether the specific information meets the predetermined standard, and deletes the storage information from the storage area when there is a command in the input information suggesting that the storage information should be deleted from the storage area. An information processing program characterized by:

Images