Information Processing Apparatus, Information Processing Method, and Program

Abstract

Using a large language model, easily display information such as dangerous locations with high accuracy on a map. 【Solution means】The information processing apparatus includes: a database information acquisition unit that acquires database information including human movement data and accident information data in the range targeted by the movement data; an input information acquisition unit that acquires input information in natural language input by a user's operation; an instruction information acquisition unit that acquires instruction information prepared in advance for input to the large language model, the instruction information including an instruction for causing the large language model to output output information conforming to a predetermined format; and a prompt generation unit that generates a prompt for input to the large language model based on the acquired database information, the input information, and the instruction information.

Description

Technical Field

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

Background Art

[0002] Conventionally, there has been a technology for displaying information on dangerous locations and the like on roads, sidewalks, etc. where people or vehicles move on a map and notifying users. For example, by displaying information on dangerous locations and the like on roads, sidewalks, etc. on a map, road construction contractors and the like can easily visually confirm locations where construction should be carried out and locations where signs such as temporary stops should be installed, and people or vehicles can pass while avoiding dangerous locations.

[0003] On the other hand, in recent years, large language models (LLMs) have been widely used in various fields. When trying to realize the above-described display on a map using a large language model, since users can give instructions in natural language, even users without specialized knowledge can visualize dangerous locations on the map. For example, Patent Document 1 discloses a technology for solving the problem that when input information obtained from a user is complex or has a vague content, the content of the answer also lacks accuracy.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] In order to find dangerous locations to be displayed on a map, a large language model can refer to pedestrian flow data, past accident information, etc. However, in order to accurately find dangerous locations and the like by referring to such information, it was necessary for the user to create input information such as an SQL statement. However, since creating this SQL statement requires specialized knowledge regarding data processing, it places a heavy burden on users who do not have specialized knowledge. Therefore, according to the prior art, there has been a problem that it is difficult for users without specialized knowledge to display information on dangerous locations and the like with high accuracy on a map using a large language model.

[0006] The present invention has been made in consideration of such circumstances, and an object thereof is to provide an information processing apparatus, an information processing method, and a program capable of easily displaying information on dangerous locations and the like with high accuracy on a map using a large language model.

Means for Solving the Problems

[0007] (1) One aspect of the present invention includes a database information acquisition unit that acquires database information including human movement data and accident information data within the range targeted by the movement data, an input information acquisition unit that acquires input information in natural language input by a user's operation, and instruction information prepared in advance for input to a large language model, the instruction information including an instruction for causing the large language model to output output information conforming to a predetermined format, and a prompt generation unit that generates a prompt for input to the large language model based on the acquired database information, the input information, and the instruction information. An explanatory text generation unit that obtains information for explaining in natural language the details of the information by inputting, into a large language model, the information obtained as a result of inputting the prompt into the large language model and instruction information prepared in advance for inputting to the large language model, the instruction information being for explaining the information in natural language. An information processing apparatus comprising the same. (2) Further, one aspect of the present invention is An information processing apparatus comprising: a database information acquisition unit that acquires database information including human movement data and accident information data in a range targeted by the movement data; an input information acquisition unit that acquires input information in natural language input by a user's operation; an instruction information acquisition unit that acquires instruction information prepared in advance for inputting to a large language model, the instruction information including an instruction for causing the large language model to output output information conforming to a predetermined format; and a prompt generation unit that generates a prompt for inputting to the large language model based on the acquired database information, the input information, and the instruction information, wherein the database information acquired by the database information acquisition unit includes at least pedestrian flow data, and the pedestrian flow data includes at least any one of the traffic volume for each movement type, the traffic volume for each age group, the traffic volume for each time, time zone, day of the week, or month. as described above. (3) Further, one aspect of the present invention is An information processing apparatus comprising: a database information acquisition unit that acquires database information including human movement data and accident information data in a range targeted by the movement data; an input information acquisition unit that acquires input information in natural language input by a user's operation; an instruction information acquisition unit that acquires instruction information prepared in advance for inputting to a large language model, the instruction information including an instruction for causing the large language model to output output information conforming to a predetermined format; and a prompt generation unit that generates a prompt for inputting to the large language model based on the acquired database information, the input information, and the instruction information, wherein the database information acquired by the database information acquisition unit includes the number of accidents or accident rate in the range targeted by the movement data. as described above. (4) Further, one aspect of the present invention is in the information processing apparatus described in (3) above, The database information acquired by the database information acquisition unit includes the number of accidents or accident rate for each movement type in the range targeted by the movement data. which is also as described above. (5) Also, one aspect of the present invention is in the information processing apparatus described above (3) and is The database information acquired by the database information acquisition unit includes the number of accident cases or accident rate for each age group within the range targeted by the movement data. as follows. (6) Also, one aspect of the present invention is in the information processing apparatus described above (3) and is The database information acquired by the database information acquisition unit includes the number of accident cases or accident rate for each time, time period, day of the week, or month within the range targeted by the movement data. as follows. (7) Also, one aspect of the present invention is in the information processing apparatus described above Any one of (1) to (3) and is The apparatus further includes an image generation unit that generates image information in which dangerous locations are indicated on a map based on information obtained as a result of inputting the prompt to the large language model. as follows. (8) Also, one aspect of the present invention is An information processing method performed by a computer, the method including: a database information acquisition step of acquiring database information including human movement data and data of accident information within the range targeted by the movement data; an input information acquisition step of acquiring input information in natural language input by a user operation; an instruction information acquisition step of acquiring instruction information prepared in advance for inputting to a large language model, the instruction information including an instruction for causing the large language model to output output information conforming to a predetermined format; a prompt generation step of generating a prompt for inputting to the large language model based on the acquired database information, the input information, and the instruction information; and an explanatory text generation step of obtaining information that explains in natural language details of the information by inputting the information obtained as a result of inputting the prompt to the large language model and instruction information prepared in advance for inputting to the large language model and for explaining the information in natural language to the large language model. as follows. (9) Also, one aspect of the present invention is An information processing method performed by a computer, comprising: a database information acquisition step of acquiring database information including human movement data and accident information data in the range targeted by the movement data; an input information acquisition step of acquiring input information in natural language input by a user's operation; an instruction information acquisition step of acquiring instruction information prepared in advance for input to a large language model, the instruction information including an instruction for outputting output information conforming to a predetermined format from the large language model; and a prompt generation step of generating a prompt for input to the large language model based on the acquired database information, the input information, and the instruction information. In the database information acquisition step, the acquired database information includes at least pedestrian flow data, and the pedestrian flow data includes at least any one of the traffic volume for each movement type, the traffic volume for each age group, the traffic volume for each time, time zone, day of the week, or month. Information processing method as follows. (10) Also, one aspect of the present invention is An information processing method performed by a computer, comprising: a database information acquisition step of acquiring database information including human movement data and accident information data in the range targeted by the movement data; an input information acquisition step of acquiring input information in natural language input by a user's operation; an instruction information acquisition step of acquiring instruction information prepared in advance for input to a large language model, the instruction information including an instruction for outputting output information conforming to a predetermined format from the large language model; and a prompt generation step of generating a prompt for input to the large language model based on the acquired database information, the input information, and the instruction information. In the database information acquisition step, the acquired database information includes the number of accidents or accident rate in the range targeted by the movement data. Information processing method as follows. (11) Also, one aspect of the present invention is On a computer the above-mentioned (8) Any one of (10) to (10) and is Program for causing a computer to execute an information processing method as follows.

Advantages of the Invention

[0008] According to the present invention, it is possible to provide an information processing apparatus, an information processing method, and a program that can easily display information on dangerous locations and the like with high accuracy on a map by using a large language model.

Brief Description of the Drawings

[0009]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Mode for Carrying Out the Invention

[0010] [Embodiment] Regarding the information processing apparatus, information processing method, and program according to aspects of the present invention, preferred embodiments will be described in detail below with reference to the accompanying drawings. Note that the aspects of the present invention are not limited to these embodiments, and also include those with various modifications or improvements. That is, the constituent elements described below include those that can be easily assumed by those skilled in the art and substantially identical ones, and the constituent elements described below can be combined as appropriate. Also, various omissions, substitutions, or changes of the constituent elements can be made without departing from the gist of the present invention. In the following drawings, in order to make each configuration easy to understand, the scale and number, etc. in each structure may be different from the scale and number, etc. in the actual structure.

[0011] [Outline of the System] FIG. 1 is a diagram showing the outline of a system according to an embodiment. First, the outline of the system 1 will be described with reference to the same figure.

[0012] Reference sign A1 is an input sentence input by the user. The user inputs the input sentence into a predetermined input range (text box) using a computer (not shown). The user assumed in the present embodiment is an ordinary person who does not have specialized knowledge regarding data processing such as the creation of SQL (Structured Query Language) sentences. Therefore, the input sentence targeted by the system 1 according to the present embodiment supports input in natural language. In an example shown in the same figure, an instruction in natural language such as "display the top 3 risks among the local roads in the target area" is described in the input sentence indicated by reference sign A1.

[0013] Note that this embodiment is not limited to an example of an input sentence consisting only of such natural language. For example, part or all of the input sentence may include information that is not natural language. Information that is not natural language included in the input sentence widely includes prompt information specified by a hash symbol or the like, a predetermined programming language, and various other forms of data.

[0014] Reference sign A2 is a pre-prepared prerequisite. The prerequisite is information regarding a request for what kind of information to output to the generation AI. That is, since the prerequisite is an instruction to the generation AI, it can also be referred to as instruction information. More specifically, the generation AI used in this embodiment may be a large language model (LLM). However, the generation AI used in this embodiment is not limited to an example of a large language model, and various generation AIs can be used.

[0015] Specifically, the prerequisite may include the following information. ·Execute a process that matches the input sentence in a dataframe ·In the dataframe, the columns hold the following Number of emergency brakes, number of passing vehicles, number of pedestrians... ·Generate a program that performs processing with SQL

[0016] As a result of inputting the input sentence indicated by reference sign A1 and the prerequisite indicated by reference sign A2 to the generation AI, SQL is obtained. It can also be said that the SQL is an instruction for generating image information corresponding to the input sentence input by the user for a predetermined application. As an example of the image information generated by a predetermined application, specifically, it is assumed that a map information with dangerous locations or the like indicated is generated. However, the image information generated based on the SQL in this embodiment is not limited to map information, and widely includes image information corresponding to the input sentence input by the user.

[0017] In the figure, an example of image information generated by a predetermined application is shown as reference sign A3. Among the reference sign A3 which is map information, reference signs H1, H2, and H3 are shown as dangerous locations. Note that the reference signs H1 to H3 correspond to the input sentence "Display the top 3 dangerous locations among the local roads in the target area" input by the user. By visually recognizing the image information of the reference sign A3, the user can easily grasp the top 3 dangerous locations among the dangerous locations in the local roads of the target area.

[0018] Note that the dangerous locations on local roads can be exemplified by, for example, locations with a narrow road width, locations with a large volume of bicycle traffic, locations with a large number of child or elderly pedestrians, locations with a high number of traffic accidents, locations with factors obstructing the view (such as overgrown vegetation), locations lacking signs, etc.

[0019] [Functional Configuration of the System] FIG. 2 is a functional configuration diagram showing an example of the functional configuration of the system according to the present embodiment. With reference to this figure, an example of the functional configuration of the system 1 for realizing the mechanism as described above will be described. In the example shown in this figure, it is assumed that the large language model (LLM) which is a generative AI is used via the cloud. Therefore, the generative AI is expressed as the LLM API. However, the LLM itself may be incorporated into the system 1.

[0020] Note that the system 1 may be realized on the cloud or on an edge terminal. In the example shown in the figure, it is assumed that the system 1 is provided as software as a service (SaaS).

[0021] First, user U creates an instruction and inputs the created instruction into computer CP. The instruction is the input sentence described with reference to FIG. 1. It is preferable that the instruction be input in natural language. As an example of the instruction, a simple one such as "Display dangerous life roads" may be used. Note that the computer CP may be a commonly used smartphone, tablet terminal, notebook computer, or the like.

[0022] The pre-prepared prompt is the prerequisite described with reference to FIG. 1.

[0023] The original data information assumes the big data stored in the original data storage unit. It is preferable that the big data include people flow data and accident data. The people flow data is data indicating the movement and stay status of people within a specific range or time. The people flow data may include people's attributes. The accident data includes information on past accidents within a specific range or time. As an example of the accident data, a bundle of information in which coordinates and the accident occurrence time are associated can be exemplified.

[0024] For the original data information, a part of the big data stored in the original data storage unit (for example, 100 pieces of data out of 10,000 pieces of data registered in one example) may be used, or if the number of specific data is small, all the data may be used. By inputting the original data information into the LLM API as described later, the LLM can determine the type of data held in the big data (for example, it can determine that the big data holds the traffic volume for each movement type), and can determine the format of the held data (for example, when holding the traffic volume, it can determine whether it is in minutes or hours). Furthermore, when making a quantitative or qualitative judgment in the LLM, the data in the big data can be referred to for judgment (for example, when there is an instruction that the traffic volume is large, it can be determined how much or more the traffic volume should be for it to be considered large).

[0025] The input text, the pre-prepared prompt, and the original data information described above are input into the LLM API. As a result of the input, SQL for the input is generated. The generated SQL undergoes the following two processes.

[0026] First, as the first process, image information is generated. In the image information generation process, first, the SQL generated by the LLM is checked. This check is a mechanical process to determine whether the SQL conforms to the format required by the application for performing the image generation process. If it is determined by this check that the format does not match, an error may be output to the user U, or a correction process of the SQL may be performed to match the format.

[0027] Next, as the second process, a description text generation process for the processing content is performed. Here, the user U may want to obtain information about what SQL statement was used to generate the image information. Since the attributes of the user U include general people who do not have specialized knowledge related to data processing such as creating SQL statements, even if the SQL statement output from the LLM is presented to the user U as it is, the content may not be understood. Therefore, in the description text generation process, the obtained SQL is used to generate a description text in natural language.

[0028] In the description text generation process, a pre-prepared prompt (a prompt in which instructions for generating a description text from SQL are described) and the obtained SQL are input into the LLM API. As a result of the input, a description text in natural language is generated.

[0029] Finally, the image information obtained in the first process and the description text obtained in the second process are presented to the user U.

[0030] Note that the above two processes may be performed simultaneously with each other, or either one of them may be performed first. When either one of them is performed first, it is preferable to perform the first process (image generation process). This is because there is a check step in the image generation process, and if an error occurs in the first process, the second process may not need to be performed.

[0031] [Functional Configuration of Information Processing Apparatus] FIG. 3 is a functional configuration diagram showing an example of the functional configuration of the information processing apparatus according to the present embodiment. With reference to this figure, an example of a specific functional configuration of the information processing apparatus 10 used to realize the system 1 will be described. Note that the information processing apparatus 10 may be downloaded to a computer CP operated by the user U, or may be executed on the cloud.

[0032] The information processing apparatus 10 includes, as its functional configuration, an acquisition unit 11, a prompt generation unit 12, a communication unit 13, a confirmation unit 14, an image generation unit 15, a description generation unit 16, and a presentation unit 17. Each of these functional units is realized, for example, using an electronic circuit. Also, each functional unit may include, as necessary, storage means such as a semiconductor memory or a magnetic hard disk device inside. Further, each function may be realized by a computer having a CPU (Central Processing Unit) and software. Also, all or part of each functional unit may be realized using hardware (for example, a circuit part; circuity) such as an ASIC (Application Specific Integrated Circuit), a PLD (Programmable Logic Device), or an FPGA (Field-Programmable Gate Array). Also, all or part of each functional unit may be realized by a combination of software and hardware.

[0033] The acquisition unit 11 acquires information used to generate information presented to the user. Specifically, the acquisition unit 11 includes a database information acquisition unit 111, an input information acquisition unit 112, and an instruction information acquisition unit 113 to acquire various types of information.

[0034] The database information acquisition unit 111 acquires database information from the database information storage unit 31. The database information includes at least movement data and accident information data. The movement data includes information related to human movement. Also, the accident data includes information related to accident information within the range targeted by the movement data.

[0035] Specifically, it is preferable that the database information includes at least flow-of-people data. Furthermore, it is preferable that the flow-of-people data includes data for each attribute of people. For example, the flow-of-people data may include information related to the traffic volume for each type of movement. The type of movement may be, for example, types such as walking, bicycle, automobile, etc. Also, the flow-of-people data may include information related to the traffic volume for each attribute of the movers, such as the age of the movers. Furthermore, the flow-of-people data may include information related to the traffic volume for each temporal range, such as time, time zone, day of the week, or month.

[0036] In this way, by storing the flow-of-people data separately for each attribute of the passers-by and each temporal range, it becomes possible to flexibly respond to the requests of the user. For example, when the user wants to know dangerous locations when moving by bicycle, if information for each bicycle already exists, it becomes possible to flexibly respond to the requests of the user. Similarly, when the user wants to know dangerous locations when moving during a specific time zone such as at night, if information for each time zone already exists, it becomes possible to flexibly respond to the requests of the user.

[0037] Specifically, it is preferable that the database information includes the number of accidents or the accident rate within the range targeted by the movement data. Furthermore, it is preferable that the database information includes the number of accidents or the accident rate for each type of movement within the range targeted by the movement data. Additionally, it is preferable that the database information includes the number of accidents or the accident rate for each age group within the range targeted by the movement data. Moreover, it is preferable that the database information includes the number of accidents or the accident rate for each time, time period, day of the week, or month within the range targeted by the movement data.

[0038] In this way, by also storing data on the number of accidents or the accident rate within the range targeted by the movement data separately for the attributes of the pedestrians and for each temporal range, it becomes possible to flexibly respond to the requests of the user.

[0039] The input information acquisition unit 112 acquires input information in natural language input by the user's operation. Specifically, the input information acquisition unit 112 acquires the input information from the computer CP operated by the user. Note that the input information is the input sentence shown in FIG. 1 and the instruction sentence shown in FIG. 2.

[0040] The instruction information acquisition unit 113 acquires the instruction information from the instruction information storage unit 32. The instruction information is information prepared in advance for input to the large language model. Also, the instruction information includes an instruction for causing the large language model to output output information conforming to a predetermined format.

[0041] The prompt generation unit 12 generates a prompt based on the information acquired by the acquisition unit 11. Specifically, the prompt generation unit 12 generates a prompt for input to the large language model based on the database information acquired by the database information acquisition unit 111, the input information acquired by the input information acquisition unit 112, and the instruction information acquired by the instruction information acquisition unit 113.

[0042] As the simplest specific method for generating a prompt, it may simply be to concatenate all the information. Further, as another method for creating a prompt, in response to the input information, the necessary database information may be searched, and a prompt may be created using only the necessary database information. Note that the method for generating a prompt according to this embodiment widely includes other methods.

[0043] The communication unit 13 performs information communication with the large language model LLM. The information communication is performed via a predetermined information communication network such as the Internet. Specifically, the communication unit 13 includes a transmission unit 131 and a reception unit 132 as functional configurations. For the sake of simplicity of explanation, the transmission unit 131 and the reception unit 132 are described as different configurations, but the transmission unit 131 and the reception unit 132 may be a single configuration.

[0044] The transmission unit 131 transmits the prompt generated by the prompt generation unit 12 to the large language model LLM. The large language model LLM performs inference using the transmitted prompt as input information and returns a result.

[0045] The reception unit 132 receives the output information output from the large language model LLM. The output information output from the large language model LLM is specifically an SQL statement. It is clear from the above-described instruction information that an SQL statement is output. For convenience, the SQL statement output from the large language model LLM is referred to as SQL1. SQL1 is provided to the confirmation unit 14 and the explanatory text generation unit 16.

[0046] The verification unit 14 verifies the SQL1 output from the large language model LLM. The verification process is a mechanical process of whether SQL1 conforms to the format required by the image generation unit 15 described later. If the verification unit 14 determines that it does not conform to the format as a result of the verification process, it may output an error to the user U. Also, if the verification unit 14 determines that it does not conform to the format as a result of the verification process, it may make it conform to the format by performing a SQL correction process. The SQL correction process may be, for example, a process of making a value within an acceptable range by correcting a value not allowed as an input value. The SQL statement after the verification process by the verification unit 14 is described as SQL2 for convenience.

[0047] The image generation unit 15 generates image information with dangerous locations indicated on a map after obtaining necessary information from the database information storage unit 31 based on SQL2. The image generation unit 15 may be an application or the like for generating image information from an SQL statement. In an example shown in the same figure, the image generation unit 15 generates image information based on SQL2, but the present embodiment is not limited to this example. For example, the verification process by the verification unit 14 may not be performed. That is, the image generation unit 15 can generate image information based on the information (SQL1 or SQL2) obtained as a result of the prompt generated by the prompt generation unit 12 being input to the large language model LLM.

[0048] The description text generation unit 16 generates information that explains the details of SQL1 output by the large language model LLM in natural language. Specifically, the description text generation unit 16 inputs the information obtained as a result of the prompt generated by the prompt generation unit 12 being input into the large language model LLM, and instruction information prepared in advance for input to the large language model LLM and used to instruct the explanation of the information in natural language, into the large language model LLM, thereby generating information that explains the details of the information in natural language. The transmission of SQL1 and the instruction information is performed by the transmission unit 131. The reception of the output result from the large language model LLM (described as the request content description text in the same figure) is performed by the reception unit 132.

[0049] The presentation unit 17 presents the image information generated by the image generation unit 15 and the processing content description text generated by the description text generation unit 16 to the user together. For example, the presentation unit 17 may present information to the computer CP from which the input information acquisition unit 112 acquires the input information.

[0050] [An example of the output by the information processing apparatus] FIG. 4 is a diagram showing an example of information presented to the user by the information processing apparatus according to the present embodiment. With reference to this figure, a detailed example of the information presented by the presentation unit 17 will be described.

[0051] FIG. 4(A) shows an example of an image generated by the image generation unit 15. The image shows map information and information specified by the user within the range shown in the map information. For example, it shows dangerous locations. The dangerous locations may be colored for each grid obtained by dividing the range shown in the map information in a predetermined pixel unit. The present embodiment is not limited to an example of being displayed in binary in this way, and other than that, it may be displayed as a heat map according to the degree of danger, or other marks indicating danger may be displayed. The mark may indicate the type of danger (not shown).

[0052] Figure 4(B) is an example of the description text generated by the description text generation unit 16. The description text includes, for example, details of the data acquired by the database information acquisition unit 111, and information such as what conditions are used to determine dangerous locations. Further, the description text may indicate the accuracy of the information, for example, by including the amount of information used in the process.

[0053] The presentation unit 17 presents the information as shown in FIGS. 4(A) and 4(B) on the display unit of the computer CP operated by the user. The information shown in FIG. 4(A) and the information shown in FIG. 4(B) may be displayed on one screen simultaneously, may be exclusively displayed by switching the screen display (for example, selection by a tab, etc.), or may be superimposed on the image information by pressing a detail button or the like to display the description text.

[0054] [Information Processing Method] FIG. 5 is a flowchart showing a series of processes of the information processing method according to the present embodiment. A series of processes of the information processing method performed using the above-described information processing apparatus 10 will be described with reference to this figure. Note that the information processing method shown in this figure is assumed to be performed by a computer.

[0055] (Step S11) First, the acquisition unit 11 acquires various information. Note that this step may be described as an acquisition step or an acquisition process. The acquisition process includes a database information acquisition process (or a database information acquisition step), an input information acquisition process (or an input information acquisition step), and an instruction information acquisition process (or an instruction information acquisition step). In the database information acquisition process, database information including human movement data and accident information data in the range targeted by the movement data is acquired. In the input information acquisition process, input information in natural language input by the user's operation is acquired. In the instruction information acquisition process, instruction information prepared in advance for input to the large language model LLM, which includes an instruction for causing the large language model LLM to output output information conforming to a predetermined format, is acquired.

[0056] (Step S12) Next, the prompt generation unit 12 generates a prompt for input to the large language model LLM based on various information (specifically, database information, input information, and instruction information) acquired by the acquisition unit 11. Note that this process may be described as a prompt generation process or a prompt generation step.

[0057] (Step S13) Next, the information processing device 10 inputs the prompt generated in Step S12 to the large language model LLM. As a result of the input, the information processing device 10 acquires SQL.

[0058] (Step S14) Next, the confirmation unit 14 checks whether the SQL acquired in Step S13 conforms to the format for performing image generation processing. Note that this process may be described as a confirmation process or a confirmation step.

[0059] (Step S15) Next, the image generation unit 15 acquires necessary information from the database information storage unit 31 based on the SQL after the confirmation in Step S14 and generates image information. Note that this process may be described as an image information generation process or an image information generation step.

[0060] (Step S16) Further, the description generation unit 16 generates a description of the SQL (which can also be referred to as the processing content) by inputting the SQL acquired in Step S13 to the large language model LLM together with predetermined instruction information. The description is in natural language and can be easily understood by the user. Note that this process may be described as a description generation process or a description generation step.

[0061] (Step S17) Finally, the presentation unit 17 presents the image information generated in Step S15 and the processing content description generated in Step S16 to the user. Note that this process may be described as a presentation process or a presentation step.

[0062] [Internal Structure] FIG. 6 is a block diagram showing an example of the internal structure of the information processing apparatus according to the present embodiment. The computer shown in the figure shows an example of a specific hardware configuration for realizing the information processing apparatus 10. The computer includes a central processing unit (processor) 901, a RAM 902, an input / output port 903, input / output devices 904 and 905, etc., and a bus 906. The computer itself can be realized using existing technologies. The central processing unit 901 executes instructions included in a program read from the RAM 902 or the like. The central processing unit 901 writes data to the RAM 902, reads data from the RAM 902, and performs arithmetic operations and logical operations according to each instruction. The RAM 902 stores data and programs. Each element included in the RAM 902 has an address and can be accessed using the address. Note that RAM is an abbreviation for "Random Access Memory". The input / output port 903 is a port for the central processing unit 901 to exchange data with external input / output devices and the like. The input / output devices 904 and 905 are input / output devices. The input / output devices 904 and 905 exchange data with the central processing unit 901 via the input / output port 903. The bus 906 is a common communication path used inside the computer. For example, the central processing unit 901 reads and writes data in the RAM 902 via the bus 906. Also, for example, the central processing unit 901 accesses the input / output port via the bus 906. Also, all or part of the information processing apparatus 10 may be realized using hardware such as an ASIC, a PLD, or an FPGA. Also, all or part of each functional unit may be realized by a combination of software and hardware.

[0063] [Summary of the Embodiment] According to the embodiments described above, the information processing apparatus 10 includes a database information acquisition unit 111, an input information acquisition unit 112, an instruction information acquisition unit 113, and a prompt generation unit 12. The database information acquisition unit 111 acquires database information including human movement data and accident information data in the range targeted by the movement data. The input information acquisition unit 112 acquires input information in natural language input by a user's operation. The instruction information acquisition unit 113 acquires instruction information prepared in advance for input to the large language model LLM, including an instruction for causing the large language model LLM to output output information conforming to a predetermined format. The prompt generation unit 12 generates a prompt for input to the large language model LLM based on the acquired database information, input information, and instruction information. By adopting such a configuration, even a user who does not have specialized knowledge regarding data processing such as the creation of SQL statements can easily display information such as dangerous locations with high accuracy on a map using the large language model LLM.

[0064] In addition, according to the above-described embodiments, it is possible to "easily display information such as dangerous locations with high accuracy on a map using a large language model". By displaying information such as dangerous locations on a map, it becomes possible to suitably maintain infrastructure such as roads. Therefore, according to the present embodiment, it is possible to contribute to Goal 9 of the Sustainable Development Goals (SDGs) led by the United Nations, "Build resilient infrastructure, promote sustainable industrialization, and foster innovation".

[0065] As described above, the embodiments of the present invention have been described in detail with reference to the drawings. However, the specific configuration is not limited to this embodiment, and design changes and the like within the scope not departing from the gist of the present invention are also included.

[0066] Also, a computer program for realizing the functions of each of the above-described apparatuses may be recorded on a computer-readable recording medium, and the program recorded on this recording medium may be read into a computer system and executed. Here, the "computer system" may include hardware such as an OS and peripheral devices. Also, the "computer-readable recording medium" refers to a flexible disk, a magneto-optical disk, a ROM, a writable non-volatile memory such as a flash memory, a portable medium such as a DVD (Digital Versatile Disc), or a storage device such as a hard disk built into a computer system.

[0067] Furthermore, the "computer-readable recording medium" also includes a volatile memory (for example, DRAM (Dynamic Random Access Memory)) inside a computer system that becomes a server or a client when a program is transmitted via a network such as the Internet or a communication line such as a telephone line, and that holds the program for a certain period of time. Also, the above program may be transmitted from a computer system storing the program in a storage device or the like to another computer system via a transmission medium or by a transmission wave in the transmission medium. Here, the "transmission medium" for transmitting the program refers to a medium having a function of transmitting information, such as a network (communication network) such as the Internet or a communication line (communication wire) such as a telephone line. Also, the above program may be for realizing a part of the above-described functions. Furthermore, it may be a so-called difference file (difference program) that can realize the above-described functions in combination with a program already recorded in a computer system.

Explanation of Signs

[0068] 1... System, 10... Information processing apparatus, 11... Acquisition unit, 111... Database information acquisition unit, 112... Input information acquisition unit, 113... Instruction information acquisition unit, 12... Prompt generation unit, 13... Communication unit, 131... Transmission unit, 132... Reception unit, 14... Confirmation unit, 15... Image generation unit, 16... Explanation text generation unit, 17... Presentation unit, 31... Database information storage unit, 32... Instruction information storage unit

Claims

1. A database information acquisition unit that acquires database information including human movement data and accident information data in the range targeted by the movement data; An input information acquisition unit that acquires input information in natural language input by a user's operation; An instruction information acquisition unit that acquires instruction information prepared in advance for input to a large language model, the instruction information including an instruction for causing the large language model to output output information conforming to a predetermined format; A prompt generation unit that generates a prompt for input to the large language model based on the acquired database information, the input information, and the instruction information; An explanatory text generation unit that inputs the information obtained as a result of inputting the prompt to the large language model and instruction information prepared in advance for input to the large language model and for explaining the information in natural language, and obtains information for explaining the details of the information in natural language; An information processing apparatus comprising the above.

2. A database information acquisition unit that acquires database information including human movement data and accident information data in the range targeted by the movement data; An input information acquisition unit that acquires input information in natural language input by a user's operation; An instruction information acquisition unit that acquires instruction information prepared in advance for input to a large language model, the instruction information including an instruction for causing the large language model to output output information conforming to a predetermined format; A prompt generation unit that generates a prompt for input to the large language model based on the acquired database information, the input information, and the instruction information; Comprising: The database information acquired by the database information acquisition unit includes at least crowd flow data; The crowd flow data includes at least one of the traffic volume for each movement type, the traffic volume for each age group, the traffic volume for each time, time zone, day of the week, or month; An information processing apparatus.

3. A database information acquisition unit that acquires database information including human movement data and accident information data in the range targeted by the movement data; An input information acquisition unit that acquires input information in natural language input by a user's operation; Instruction information prepared in advance for input to a large language model, including instruction information for obtaining instruction information including an instruction for causing the large language model to output output information conforming to a predetermined format A prompt generation unit that generates a prompt for input to the large language model based on the acquired database information, the input information, and the instruction information Comprising The database information acquired by the database information acquisition unit includes the number of accident cases or accident rate in the range targeted by the movement data Information processing apparatus

4. The database information acquired by the database information acquisition unit includes the number of accident cases or accident rate for each type of movement in the range targeted by the movement data The information processing apparatus according to claim 3

5. The database information acquired by the database information acquisition unit includes the number of accident cases or accident rate for each age group in the range targeted by the movement data The information processing apparatus according to claim 3

6. The database information acquired by the database information acquisition unit includes the number of accident cases or accident rate for each time, time zone, day of the week, or month in the range targeted by the movement data The information processing apparatus according to claim 3

7. Further comprising an image generation unit that generates image information in which dangerous locations are shown on a map based on the information obtained as a result of inputting the prompt to the large language model The information processing apparatus according to any one of claims 1 to 3

8. An information processing method performed by a computer, comprising A database information acquisition step of acquiring database information including human movement data and data on accident information in the range targeted by the movement data An input information acquisition step of acquiring input information in natural language input by a user's operation An instruction information acquisition step of acquiring instruction information prepared in advance for input to a large language model, including an instruction for causing the large language model to output output information conforming to a predetermined format A prompt generation step of generating a prompt for input to the large language model based on the acquired database information, the input information, and the instruction information An explanatory text generation step of obtaining information for explaining in natural language the details of the information by inputting, into a large language model, information obtained as a result of inputting the prompt into the large language model and instruction information prepared in advance for inputting to the large language model and for explaining the information in natural language. An information processing method having the above. **Claim 9**: An information processing method performed by a computer, A database information acquisition step of acquiring database information including human movement data and accident information data in the range targeted by the movement data. An input information acquisition step of acquiring input information in natural language input by a user's operation. An instruction information acquisition step of acquiring instruction information prepared in advance for inputting to a large language model, the instruction information including an instruction for causing the large language model to output output information conforming to a predetermined format. A prompt generation step of generating a prompt for inputting to the large language model based on the acquired database information, the input information, and the instruction information. Comprising: The database information acquired in the database information acquisition step includes at least people flow data. The people flow data includes at least any one of the traffic volume for each movement type, the traffic volume for each age group, the traffic volume for each time, time period, day of the week, or month. An information processing method having the above. **Claim 10**: An information processing method performed by a computer, A database information acquisition step of acquiring database information including human movement data and accident information data in the range targeted by the movement data. An input information acquisition step of acquiring input information in natural language input by a user's operation. An instruction information acquisition step of acquiring instruction information prepared in advance for inputting to a large language model, the instruction information including an instruction for causing the large language model to output output information conforming to a predetermined format. A prompt generation step of generating a prompt for inputting to the large language model based on the acquired database information, the input information, and the instruction information. Comprising: The database information acquired in the database information acquisition step includes the number of accidents or accident rate in the range targeted by the movement data. An information processing method having the above. A program for causing a computer to execute the information processing method according to any one of claims 8 to 10.

Images