System, Program, and Method for Manufacturing Document Information

Abstract

An object of the present invention is to provide a system suitable for generating a text related to travel. 【Means for Solving the Problem】The system according to the present invention includes one or more computers. The first text is a text related to travel and includes a plurality of second texts. The control circuit of the one or more computers acquires definition book information indicating the direction of the first text, and based on the definition book information, gives an instruction to the first LLM to generate configuration information indicating the direction of the plurality of second texts, and based on the configuration information and the definition book information generated by the first LLM, gives an instruction to the second LLM to generate a plurality of second text information indicating the plurality of second texts.

Description

Technical Field

[0001] The present invention relates to a system, a program, and a method for producing text information.

Background Art

[0002] As an invention related to a conventional system, for example, a picture book generation system described in Patent Document 1 is known. This picture book generation system includes a reception unit, an illustration unit, a story unit, and an integration unit. The reception unit receives user input information including at least one of a photograph, text, and a theme. The illustration unit transmits an illustration generation instruction to an illustration generation means using the user input information. The story unit transmits a story generation instruction to a story generation means using the user input information. The integration unit integrates the illustration generated by the illustration generation means and the story generated by the story generation means to generate picture book data. According to such a picture book generation system, a picture book having an illustration and a story suitable for the illustration can be generated using the photograph taken by the user and the information input by the user.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] By the way, a system suitable for generating a text related to travel is required.

[0005] Therefore, an object of the present invention is to provide a system, a program, and a method for producing text information suitable for generating a text related to travel.

Means for Solving the Problems

[0006] The first aspect is The system includes one or more computers, The first article is an article related to travel and includes a plurality of second articles, The control circuit of the one or more computers obtains definition book information indicating the policy of the first article, gives an instruction to generate configuration information indicating the policy of the plurality of second articles to a first LLM (Large Language Models) based on the definition book information, gives an instruction to generate a plurality of second article information indicating the plurality of second articles to a second LLM based on the configuration information and the definition book information generated by the first LLM, obtains the plurality of second article information generated by the second LLM, It is a system.

[0007] The second form is The control circuit of the one or more computers obtains theme information indicating the theme of the first article, gives an instruction to generate configuration information indicating the policy of the plurality of second articles to a first LLM based on the definition book information and the theme information, It is the system described in the first form.

[0008] The third form is The control circuit of the one or more computers generates first article information indicating the first article by combining the plurality of second articles, It is the system described in the first form or the second form.

[0009] The fourth form is The control circuit of the one or more computers obtains the project plan information input by the user, obtains the definition book information by converting the project plan information into a data exchange format by a third LLM, It is the system described in the second form.

[0010] The fifth form is The first text is a travel article, The theme information includes the area input by the user. It is the system described in the second form.

[0011] The sixth form is The first text is a travel article, The theme information includes the area input by the user, The proposal information includes a proposal for generating a travel article related to the area included in the theme information. It is the system described in the fifth form.

[0012] The seventh form is The first text includes a plurality of items, The plurality of second texts correspond to the plurality of items. It is the system described in the fourth form or the sixth form.

[0013] The eighth form is The proposal information includes a plurality of paragraphs, The control circuit of the one or more computers is Based on the plurality of paragraphs included in the proposal information, cause the third LLM to generate the definition book information including the plurality of items. It is the system described in the seventh form.

[0014] The ninth form is The control circuit of the one or more computers is Based on the definition book information and the theme information, collect the first collection information using the fourth LLM, Based on the definition book information, the theme information, and the first collection information, give an instruction to cause the first LLM to generate the configuration information. It is the system described in any one of the second form, the fourth form to the eighth form.

[0015] The tenth form is The first article is a travel article, The definition book information includes instructions for generating articles related to the area specified by the theme information and instructions for creating articles related to a plurality of spots included in the area specified by the theme information, In the control circuit of the one or more computers, Collect the first collection information related to the area specified by the theme information and the plurality of spots using the fourth LLM, The system according to the ninth aspect.

[0016] The eleventh aspect is, The control circuit of the one or more computers, Based on the definition book information, the theme information, and the configuration information, collect second collection information using the fifth LLM, Based on the configuration information and the second collection information, give an instruction to cause the second LLM to generate the plurality of second article information, The system according to any one of the second aspect, the fourth aspect to the tenth aspect.

[0017] The twelfth aspect is, The second collection information includes external information existing on the Internet and internal information stored in a server whose access from the outside is restricted, The internal information includes the first article generated in the past, The system according to the eleventh aspect.

[0018] The thirteenth aspect is, The control circuit of the one or more computers, When collecting the second collection information using the fifth LLM, do not collect information that does not meet a predetermined condition as the second collection information, The system according to the twelfth aspect.

[0019] The fourteenth aspect is, The control circuit of the one or more computers, Using the sixth LLM, perform fact-checking on the plurality of second text information. The system according to any one of the first to thirteenth aspects.

[0020] The fifteenth aspect is The first text is a travel article, At least a part of the plurality of second text information includes a link to information related to travel products. The system according to any one of the first to fourteenth aspects.

[0021] The sixteenth aspect is The configuration information includes an instruction for causing the second LLM to output one or more image information related to the plurality of second texts, At least a part of the plurality of second text information includes the one or more image information. The system according to any one of the first to fifteenth aspects.

[0022] The seventeenth aspect is The configuration information includes an instruction for causing the second LLM to output the URLs referred to during the generation of the plurality of second texts, At least a part of the plurality of second text information includes the URLs referred to by the second LLM during the generation of the plurality of second texts. The system according to any one of the first to sixteenth aspects.

[0023] The eighteenth aspect is A program, The first text is a text related to travel and includes a plurality of second texts, The program causes a control circuit of one or more computers to Obtain definition book information indicating the policy of the first text, Based on the definition book information, issue an instruction for causing the first LLM (Large Language Models) to generate configuration information indicating the policy of the plurality of second texts. Based on the configuration information and the definition document information generated by the first LLM, an instruction is given to cause the second LLM to generate a plurality of second text information indicating the plurality of second texts. To obtain the plurality of second text information generated by the second LLM. It is a program.

[0024] The nineteenth form is A method for manufacturing text information, The first text is a text related to travel and includes a plurality of second texts. The method for manufacturing text information causes a control circuit of one or more computers to Obtain definition document information indicating the policy of the first text. Based on the definition document information, an instruction is given to cause the first LLM (Large Language Models) to generate configuration information indicating the policies of the plurality of second texts. Based on the configuration information and the definition document information generated by the first LLM, an instruction is given to cause the second LLM to generate a plurality of second text information indicating the plurality of second texts. To obtain the plurality of second text information generated by the second LLM. It is a method for manufacturing text information.

Effect of the Invention

[0025] According to the present invention, a system, a program, and a method for manufacturing text information suitable for generating a text related to travel can be provided.

Brief Description of the Drawings

[0026]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

DETAILED DESCRIPTION OF THE INVENTION

[0027] (Embodiment) A system 1, a program PG, and a text information manufacturing method according to an embodiment of the present invention will be described with reference to the drawings.

[0028] [Structure of the System] First, the overall configuration of the system 1 will be described with reference to the drawings. FIG. 1 is a block diagram of the system 1. FIG. 2 is a block diagram of the first terminal 10. FIG. 3 is a block diagram of the first server 210. FIG. 4 is a block diagram of the second server 310.

[0029] The system 1 shown in FIG. 1 includes a first terminal 10, a first server 210, and a second server 310. The first terminal 10, the first server 210, and the second server 310 can communicate with each other via a communication network. The network is the Internet, an intranet, or the like.

[0030] The first terminal 10 is a computer used by a user. The user is, for example, a staff member of a travel company who wants to generate a travel product title. The first terminal 10 is, for example, a personal computer, a smartphone, or a tablet terminal. In the present embodiment, the first terminal 10 is a personal computer arranged in a travel company. As shown in FIG. 2, the first terminal 10 includes a control circuit 12, a memory circuit 14, a network interface 16, a graphic processing unit 18, a display 20, an operation unit 26, and a touch panel 28.

[0031] The memory circuit 14 stores programs and data. The memory circuit 14 is, for example, a combination of a ROM (Read Only Memory), a RAM (Random Access Memory), and a storage (e.g., flash memory or hard disk).

[0032] The program includes, for example, the following programs. · Program of an OS (Operating System) · Program of an application that executes information processing (e.g., a web browser or the target application described later)

[0033] The data includes, for example, the following data. · Database referred to in information processing · Data obtained by executing information processing (i.e., the execution result of information processing)

[0034] The control circuit 12 realizes the functions of the first terminal 10 by executing the programs stored in the memory circuit 14. The control circuit 12 is, for example, at least one of the following. · CPU (Central Processing Unit) · GPU (Graphic Processing Unit) · ASIC (Application Specific Integrated Circuit) ·FPGA (Field Programmable Array)

[0035] The network interface 16 controls the communication between the first terminal 10 and an external device. The external device is the first server 210.

[0036] The graphic processing unit 18 causes the display 20 to display an image based on the video data generated by the control circuit 12. The display 20 is a liquid crystal display or an organic EL (Electro Luminescence) display.

[0037] The operation unit 26 generates an operation signal based on a user operation by the touch panel 28 and outputs the operation signal to the control circuit 12.

[0038] The first server 210 is a computer. The first server 210 is an application server that operates according to a predefined logic based on a request from a user. The first server 210 receives an instruction from the first terminal 10 and outputs a processing instruction to the second server 310. The first server 210 is, for example, a server of Dify (registered trademark). As shown in FIG. 3, the first server 210 includes a control circuit 212, a storage circuit 214, and a network interface 216.

[0039] The storage circuit 214 stores programs and data. The storage circuit 214 is, for example, a combination of a ROM (Read Only Memory), a RAM (Random Access Memory), and a storage (e.g., flash memory or hard disk).

[0040] The control circuit 212 realizes the functions of the first server 210 by executing the programs stored in the storage circuit 214. The control circuit 212 is, for example, at least one of the following. ·CPU (Central Processing Unit) ·GPU (Graphic Processing Unit) ·ASIC (Application Specific Integrated Circuit) ·FPGA (Field Programmable Array)

[0041] The control circuit 212 includes, as functional blocks, a definition document information acquisition means 220, a configuration information generation instruction means 222, a second text information generation instruction means 224, a theme information acquisition means 226, a fact check means 228, a second text information acquisition means 230, a proposal document information acquisition means 232, a first collected information collection means 234, a first text information generation means 236, and a second collected information collection means 238.

[0042] The network interface 216 controls communication between the first server 210 and an external device. The external devices are the first terminal 10 and the second server 310.

[0043] The second server 310 is an inference server equipped with an LLM (Large Language Model). As shown in FIG. 4, the second server 310 includes a control circuit 312, a memory circuit 314, and a network interface 316.

[0044] The memory circuit 314 stores programs and data. The memory circuit 314 is, for example, a combination of a ROM (Read Only Memory), a RAM (Random Access Memory), and a storage (e.g., flash memory or hard disk).

[0045] The control circuit 312 realizes the functions of the second server 310 by executing the programs stored in the memory circuit 314. The control circuit 312 is, for example, at least one of the following. ·CPU (Central Processing Unit) ·GPU (Graphic Processing Unit) · ASIC (Application Specific Integrated Circuit) · FPGA (Field Programmable Array)

[0046] The network interface 316 controls communication between the second server 310 and an external device. The external device is the first server 210.

[0047] Also, the second server 310 includes a first LLM 318, a second LLM 320, a third LLM 322, a fourth LLM 324, a fifth LLM 326, and a sixth LLM 328. The LLM is a natural language processing model trained using a large amount of text data. The LLM can perform natural language processing such as text generation, for example, by fine-tuning a pre-trained learned model.

[0048] [Operation of System 1] The operation of System 1 will be described below with reference to the drawings. FIG. 5 is a diagram showing the first text information I1. FIG. 6 is a flowchart showing the operation of System 1. FIGS. 7 and 8 are diagrams showing the proposal information IK. FIG. 9 is a diagram showing the first definition information IS1. FIG. 10 is a diagram showing the second definition information IS2. FIG. 11 is a diagram showing the first prompt P1. FIG. 12 is a diagram showing the configuration information IST. FIG. 13 is a diagram showing the second prompt P2. FIG. 14 is a diagram showing the second text information I2-1, I2-2.

[0049] By having the control circuit 212 of the first server 210 read the program PG stored in the storage circuit 214, these programs PG cause the control circuit 212 of the first server 210 to execute the operations described below. The program PG includes a definition document information acquisition means 220, a configuration information generation instruction means 222, a second text information generation instruction means 224, a theme information acquisition means 226, a fact check means 228, a second text information acquisition means 230, a project document information acquisition means 232, a first collection information collection means 234, a first text information generation means 236, and a second collection information collection means 238, and causes the control circuit 212 to execute them.

[0050] Also, the storage circuit 214 of the first server 210 stores a first prompt P1, a second prompt P2, and a third prompt P3. The first prompt P1, the second prompt P2, and the third prompt P3 may be generated by the user or by the LLM.

[0051] The system 1 aims to generate the first text information I1 shown in FIG. 5. The first text information I1 is a digital document. In this embodiment, the first text information I1 shown in FIG. 5 is a web page displayed on the website of a travel company. The first text information I1 includes a first text Tx1 related to travel. Specifically, the first text Tx1 related to travel is a travel article introducing 10 recommended hotels in Australia. The first text Tx1 includes a plurality of items S1 to SN. In FIG. 5, only items S1 and S2 are described. N is an integer of 2 or more. The items S1 to SN are the titles of the topics of the first text Tx1. That is, the items S1 to SN are the titles for segmenting and explaining the first text Tx1.

[0052] Also, the first text Tx1 includes second texts Tx2-1 to Tx2-N. The second texts Tx2-1 to Tx2-N correspond to the items S1 to SN. The second texts Tx2-1 to Tx2-N are described below each of the items S1 to SN. Each of the second texts Tx2-1 to Tx2-N is a text explaining the details of the items S1 to SN.

[0053] System 1 generates the first text Tx1 according to the following procedure. First, System 1 causes the first LLM 318 to generate the composition of the first text Tx1. The composition of the first text Tx1 is the generation of items S1 to SN. Then, System 1 causes the second LLM 320 to generate second texts Tx2-1 to Tx2-N corresponding to each of the items S1 to SN that are the composition of the first text Tx1. Finally, System 1 generates the first text Tx1 by combining the second texts Tx2-1 to Tx2-N.

[0054] The details of the operation of System 1 will be described below with reference to FIG. 6. First, the user inputs the planning document information IK shown in FIGS. 7 and 8 by operating the operation unit 26 of the first terminal 10. Thereby, the control circuit 12 of the first terminal 10 acquires the planning document information IK (step S1).

[0055] The planning document information IK includes a text created by the user. The planning document information IK includes a planning document for generating a travel article regarding an area included in the theme information ITE described later. In the present embodiment, the planning document information IK includes a planning document for creating a first text Tx1 (travel article) of "10 recommended hotels in ○○ (country name)". The theme information ITE described later includes an area corresponding to ○○ (country name). The planning document information IK includes a writing procedure A1, a published article A2, and an outline A3 of the first text Tx1 (travel article).

[0056] The writing procedure A1 shows the creation procedure of the first text Tx1 (travel article) and includes three steps in the present embodiment. Specifically, keyword research is performed in step 1, detailed research on hotels is performed using the keywords in step 2, and a rough composition of the first text Tx1 is created in step 3.

[0057] The published article A2 is the URL of a published travel article. The published article A2 is the survey result by the user.

[0058] Summary of the first article Tx1: A3 is used for generating the first article Tx1. The summary of the first article Tx1 includes a plurality of paragraphs s1 to sM such as "eye-catching" and "description sheet". M is an integer of 2 or more. Each of the plurality of paragraphs s1 to sM includes the conditions required for <the main text> which is a part of the travel article included in the first article Tx1. Each of the plurality of paragraphs s1 to sM also includes <an example of the main text> corresponding to <the main text>. <An example of the main text> may be an article created by the user or an article created by the LLM. In this way, the proposal information IK includes information regarding a plurality of items S1 to SN included in the first article Tx1. Note that the plurality of paragraphs s1 to sM included in the proposal information IK may or may not match the plurality of items S1 to SN. In FIG. 8, the plurality of paragraphs s1 to sM included in the proposal information IK do not match the plurality of items S1 to SN.

[0059] Next, the user inputs the theme information ITE by operating the operation unit 26 of the first terminal 10. Thereby, the control circuit 12 of the first terminal 10 acquires the theme information ITE (step S2). The theme information ITE indicates the theme of the first article Tx1. In the present embodiment, the theme information ITE includes an area input by the user. The area means a region, and in the present embodiment, it means a country. In the present embodiment, the user inputs "Australia" as the area of the theme information ITE. In this way, the proposal information IK includes information for generating a travel article without specifying an area. On the other hand, the theme information ITE includes an area. Therefore, by combining the proposal information IK and the theme information ITE, the input information for generating a travel article in a specific area is complete.

[0060] Next, the control circuit 12 of the first terminal 10 transmits the project plan information IK and the theme information ITE to the first server 210 via the network interface 16. Accordingly, the network interface 216 of the first server 210 receives the project plan information IK and the theme information ITE, and outputs the project plan information IK and the theme information ITE to the control circuit 212. Thereby, the control circuit 212 (control circuit of one or more computers, project plan information acquisition means 232, theme information acquisition means 226) of the first server 210 acquires the project plan information IK and the theme information ITE input by the user (step S11).

[0061] Next, the control circuit 212 of the first server 210 generates the first input information II1 to be input to the third LLM 322 (step S12). The first input information II1 includes the third prompt P3, the project plan information IK, and the theme information ITE. The third prompt P3 is stored by the storage circuit 214. The third prompt P3 is an instruction for causing the third LLM 322 to generate the first definition book information IS1 and the second definition book information IS2 based on the project plan information IK and the theme information ITE.

[0062] The first definition book information IS1 and the second definition book information IS2 (definition book information) indicate the direction of the first text Tx1. Specifically, as shown in FIG. 9, the first definition book information IS1 includes instructions for generating a plurality of paragraphs s1 to sM included in the project plan information IK, and example sentences of the texts included in the plurality of paragraphs s1 to sM. These instructions are generated by the third LLM 322 based on the writing procedure A1 of the project plan information IK, the published article A2, and the outline A3 of the first text Tx1 (travel article). Further, as shown in FIG. 10, the second definition book information IS2 includes a plurality of items S1 to SN. The plurality of items S1 to SN are generated by the third LLM 322 based on the writing procedure A1 of the project plan information IK, the published article A2, and the outline A3 of the first text Tx1 (travel article).

[0063] The above-described first definition document information IS1 and second definition document information IS2 are described in a data conversion format. In this embodiment, the first definition document information IS1 and the second definition document information IS2 are described in XML (eXtensible Markup Language) format. Therefore, as shown in FIG. 11, the third prompt P3 includes the following three instructions. · Generate the first definition document information IS1 by converting the project plan information IK into XML format. · Generate the second definition document information IS2 including a plurality of items S1 to SN by extracting a plurality of paragraphs s1 to sM from the project plan information IK. · Describe the second definition document information IS2 in XML format.

[0064] The control circuit 212 of the first server 210 transmits the first input information II1 to the second server 310 via the network interface 216 (step S13). Accordingly, the network interface 316 of the second server 310 receives the first input information II1 and outputs the first input information II1 to the control circuit 312. Thereby, the control circuit 312 of the second server 310 acquires the first input information II1.

[0065] The third LLM 322 reads the project plan information IK and the theme information ITE according to the instructions of the third prompt P3. The third LLM 322 generates the first definition document information IS1 and the second definition document information IS2 based on the project plan information IK and the theme information ITE according to the instructions of the third prompt P3.

[0066] Next, the control circuit 312 of the second server 320 transmits the first definition book information IS1 and the second definition book information IS2 to the first server 210 via the network interface 316. In response, the network interface 216 of the first server 210 receives the first definition book information IS1 and the second definition book information IS2, and outputs the first definition book information IS1 and the second definition book information IS2 to the control circuit 212. Thereby, the control circuit 212 of the first server 210 acquires the first definition book information IS1 and the second definition book information IS2 (step S14). As described above, the control circuit 212 (the control circuit of one or more computers, the definition book information acquisition means 220) of the first server 210 causes the third LLM 322 to convert the project plan information IK into the XML format (data exchange format) in steps S12 to S14, thereby acquiring the first definition book information IS1 and the second definition book information IS2 (definition book information). At this time, the control circuit 212 (the control circuit of one or more computers, the definition book information acquisition means 220) of the first server 210 causes the third LLM 322 to generate the second definition book information IS2 (definition book information) including a plurality of items S1 to SN based on the plurality of paragraphs s1 to sM included in the project plan information IK.

[0067] Next, the control circuit 212 (the control circuit of one or more computers, the first collection information collection means 234) of the first server 210 collects the first collection information IR1 using the fourth LLM 324 based on the first definition book information IS1, the second definition book information IS2 (definition book information), and the theme information ITE (step S15). The control circuit 212 of the first server 210 transmits the first definition book information IS1, the second definition book information IS2, and the theme information ITE to the second server 310 via the network interface 216. The network interface 316 of the second server 310 receives the first definition book information IS1, the second definition book information IS2, and the theme information ITE, and outputs the first definition book information IS1, the second definition book information IS2, and the theme information ITE to the control circuit 312. Thereby, the control circuit 312 of the second server 310 acquires the first definition book information IS1, the second definition book information IS2, and the theme information ITE.

[0068] Here, the project plan information IK includes instructions for generating text related to the area specified by the theme information ITE and instructions for creating text related to a plurality of spots included in the area specified by the theme information ITE. A spot refers to a hotel or surrounding facilities of a hotel, etc. Specifically, the project plan information IK includes the writing procedure A1. The writing procedure A1 includes three steps. Specifically, keyword research is carried out in step 1, a detailed investigation of the hotel is carried out using the keywords in step 2, and a rough structure of the first text Tx1 is created in step 3. Since the first definition book information IS1 is the information obtained by converting the project plan information IK into XML format, it includes information corresponding to the above writing procedure A1. Therefore, the first definition book information IS1 also includes instructions for generating text related to the area specified by the theme information ITE and instructions for creating text related to a plurality of spots included in the area specified by the theme information ITE. Thus, the fourth LLM324 generates the first collected information IR1 by executing step 1 and step 2. The first collected information IR1 is information related to the first text Tx1 and is used for generating the first text Tx1. The first collected information IR1 includes, for example, the URL of the hotel and the URL of the surrounding facilities of the hotel, etc.

[0069] The control circuit 312 of the second server 310 transmits the first collected information IR1 to the first server 210 through the network interface 316. Accordingly, the network interface 216 of the first server 210 receives the first collected information IR1 and outputs the first collected information IR1 to the control circuit 212. Thereby, the control circuit 212 of the first server 210 obtains the first collected information IR1. As described above, in step S15, the control circuit 212 (control circuit of one or more computers, first collected information collection means 234) of the first server 210 collects the first collected information IR1 related to the area and a plurality of spots specified by the theme information ITE using the fourth LLM324.

[0070] Next, the control circuit 212 of the first server 210 generates second input information II2 to be input to the first LLM 318 (step S16). The second input information II2 includes the first prompt P1, theme information ITE, first definition book information IS1, second definition book information IS2, and first collection information IR1 shown in FIG. 11. The first prompt P1 is an instruction for causing the first LLM 318 to generate configuration information IST shown in FIG. 12 based on the theme information ITE, the first definition book information IS1, the second definition book information IS2, and the first collection information IR1. More specifically, the configuration information IST indicates the guidelines of a plurality of second sentences Tx2-1 to Tx2-N included in each of the plurality of items S1 to SN. In the present embodiment, the configuration information IST includes titles given to the plurality of items S1 to SN and instructions for creating the main text of each of the plurality of items S1 to SN. Further, the configuration information IST includes an instruction for causing the second LLM 320 to output one or more pieces of image information IP related to the plurality of second sentences Tx2-1 to Tx2-N. The one or more pieces of image information IP include images included in the web page when the first sentence Tx1 is a web page of a travel company's website. Furthermore, the configuration information IST includes an instruction for causing the second LLM 320 to output the URLs referred to when generating the plurality of second sentences Tx2-1 to Tx2-N.

[0071] So that the configuration information IST as described above is generated, the first prompt P1 includes a task, survey results, and instructions. The task describes an instruction to generate an outline (structure) of a travel article (the first sentence Tx1). The survey results describe an instruction to analyze the first collection information IR1. Also, the instructions describe an instruction to generate an outline (structure) of a travel article (the first sentence Tx1) using the first collection information IR1. Further, an instruction to output one or more pieces of image information IP related to the plurality of second sentences Tx2-1 to Tx2-N is described. The instructions describe an instruction to output the URLs referred to when generating the plurality of second sentences Tx2-1 to Tx2-N. Also, the instructions describe an instruction to refer to the second definition book information IS2 and an example of the configuration information IST.

[0072] Next, the control circuit 212 of the first server 210 transmits the second input information II2 to the second server 310 via the network interface 216 (step S17). In response, the network interface 316 of the second server 310 receives the second input information II2 and outputs the second input information II2 to the control circuit 312. Thereby, the control circuit 312 of the second server 310 acquires the second input information II2.

[0073] The first LLM 318 analyzes the first collection information IR1 according to the instruction of the first prompt P1. The first LLM 318 generates the configuration information IST based on the analysis results of the first definition book information IS1, the second definition book information IS2, the theme information ITE, and the first collection information IR1. At this time, as shown in FIG. 12, the first LLM 318 includes in the configuration information IST an instruction for causing the second LLM 320 to output the URLs referred to when generating the plurality of second sentences Tx2-1 to Tx2-N, and one or more pieces of image information IP related to the plurality of second sentences Tx2-1 to Tx2-N. As described above, in steps S16 and S17, the control circuit 212 (the control circuit of one or more computers, the configuration information generation instruction means 222) of the first server 210 is based on the theme information ITE, the first collection information IR1, the first definition book information IS1, and the second definition book information IS2 (definition book information), and gives an instruction to the first LLM 318 to generate the configuration information IST indicating the direction of the plurality of second sentences Tx2-1 to Tx2-N included in each of the plurality of items S1 to SN.

[0074] Next, the control circuit 312 of the second server 310 transmits the configuration information IST to the first server 210 via the network interface 316. In response, the network interface 216 of the first server 210 receives the configuration information IST and outputs the configuration information IST to the control circuit 212. Thereby, the control circuit 212 of the first server 210 acquires the configuration information IST (step S18).

[0075] Next, the control circuit 212 of the first server 210 generates third input information II3 to be input to the second LLM 320 (step S19). The third input information II3 includes the second prompt P2, theme information ITE, first definition book information IS1, second definition book information IS2, first collection information IR1, and configuration information IST shown in FIG. 13. The second prompt P2 is an instruction for causing the second LLM 320 to generate second sentence information I2-1 to I2-N indicating a plurality of second sentences Tx2-1 to Tx2-N included in each of the plurality of items S1 to SN based on the configuration information IST, the first definition book information IS1, and the second definition book information IS2 (definition book information) generated by the first LLM 318. In the present embodiment, as shown in FIG. 13, the second prompt P2 includes an instruction for causing the second LLM 320 to read the configuration information IST, and an instruction for causing the second LLM 320 to generate a plurality of second sentences Tx2-1 to Tx2-N with reference to the first collection information IR1.

[0076] Next, the control circuit 212 of the first server 210 transmits the third input information II3 to the second server 310 via the network interface 216 (step S20). In response, the network interface 316 of the second server 310 receives the third input information II3 and outputs the second input information II2 to the control circuit 312. Thereby, the control circuit 312 of the second server 310 acquires the third input information II3.

[0077] The second LLM 320 generates a plurality of second sentences Tx2-1 to Tx2-N shown in FIG. 14 according to the instructions of the second prompt P2. Specifically, as shown in FIG. 14, the second LLM 320 creates the text of the lead sentence of the second sentence Tx2-1 according to the instructions of the lead sentence of the configuration information IST. Similarly, the second LLM 320 creates the text of the hotel overview of the second sentence Tx2-1 according to the instructions of the hotel overview of the configuration information IST. However, the configuration information IST includes instructions for causing the second LLM 320 to output the URLs referred to during the generation of the plurality of second sentences Tx2-1 to Tx2-N, and one or more pieces of image information IP related to the plurality of second sentences Tx2-1 to Tx2-N. Therefore, the second LLM 320 searches for one or more pieces of image information IP and attaches the image shown by the image information IP near the text of the lead sentence of the second sentence Tx2-1. Further, the second LLM 320 describes, for example, the hotel URL in the text of the hotel overview of the second sentence Tx2-1. In this way, at least a part of the plurality of second sentence information I2-1 to I2-N includes one or more pieces of image information IP and the URLs referred to by the second LLM 320 during the generation of the plurality of second sentences Tx2-1 to Tx2-N.

[0078] As described above, the control circuit 212 of the first server 210 (the control circuit of one or more computers, the second sentence information generation instruction means 224) generates, in steps S19 and S20, a plurality of second sentence information I2-1 to I2-N indicating a plurality of second sentences Tx2-1 to Tx2-N included in each of the plurality of items S1 to SN based on the configuration information IST, the first definition book information IS1, and the second definition book information IS2 (definition book information) generated by the first LLM 318. An instruction for causing the second LLM 320 to perform is given.

[0079] Next, the control circuit 312 of the second server 310 transmits a plurality of second text information I2-1 to I2-N to the first server 210 via the network interface 316. In response, the network interface 216 of the first server 210 receives the plurality of second text information I2-1 to I2-N and outputs the plurality of second text information I2-1 to I2-N to the control circuit 212. Thereby, the control circuit 212 (control circuit of one or more computers, second text information acquisition means 230) of the first server 210 acquires the plurality of second text information I2-1 to I2-N generated by the second LLM 320 (step S21).

[0080] Next, the control circuit 212 (control circuit of one or more computers, fact-checking means 228) of the first server 210 performs fact-checking on the plurality of second texts Tx2-1 to Tx2-N using the sixth LLM 328 (step S22). More specifically, the storage circuit 214 of the first server 210 stores a prompt for fact-checking. The control circuit 212 of the first server 210 transmits the prompt for fact-checking and the plurality of second text information I2-1 to I2-N to the second server 310 via the network interface 216. In response, the network interface 316 of the second server 310 receives the prompt for fact-checking and the plurality of second text information I2-1 to I2-N, and outputs the prompt for fact-checking and the plurality of second text information I2-1 to I2-N to the control circuit 312. Thereby, the control circuit 312 of the second server 310 acquires the prompt for fact-checking and the plurality of second text information I2-1 to I2-N.

[0081] The 6th LLM 328 performs fact-checking on a plurality of second sentences Tx2-1 to Tx2-N in accordance with the instructions of the prompt for fact-checking. Specifically, the 6th LLM 328 collects information regarding the plurality of second sentences Tx2-1 to Tx2-N via a network, and compares the collected information with the plurality of second sentences Tx2-1 to Tx2-N. Then, the 6th LLM 328 determines whether the plurality of second sentences Tx2-1 to Tx2-N contain errors. The 6th LLM 328 determines, for example, whether numerical values such as the population of Australia are incorrect, or whether the address of a hotel is incorrect. The 6th LLM 328 generates fact-checking information IFC indicating the result of the fact-checking. Therefore, the fact-checking information IFC includes information pointing out errors contained in the plurality of second sentences Tx2-1 to Tx2-N. However, the fact-checking information IFC may indicate that the plurality of second sentences Tx2-1 to Tx2-N do not contain errors when the plurality of second sentences Tx2-1 to Tx2-N do not contain errors.

[0082] Next, the control circuit 312 of the second server 310 transmits the fact-checking information IFC to the first server 210 via the network interface 316. Accordingly, the network interface 216 of the first server 210 receives the fact-checking information IFC and outputs the fact-checking information IFC to the control circuit 212. Thereby, the control circuit 212 of the first server 210 acquires the fact-checking information IFC.

[0083] Next, the control circuit 212 (control circuit of one or more computers, first sentence information generation means 326) of the first server 210 generates first sentence information I1 indicating the first sentence Tx1 by combining the plurality of second sentences Tx2-1 to Tx2-N (step S23). As shown in FIG. 5, the first sentence Tx1 is a web page included in the website of a travel agency.

[0084] Next, the control circuit 212 of the first server 210 transmits the first text information I1 and the fact-check information IFC to the first terminal 10 via the network interface 216. Accordingly, the network interface 16 of the first terminal 10 receives the first text information I1 and the fact-check information IFC, and outputs the first text information I1 and the fact-check information IFC to the control circuit 12. Thereby, the control circuit 12 of the first terminal 10 acquires the first text information I1 and the fact-check information IFC (step S4).

[0085] Finally, the control circuit 12 of the first terminal 10 causes the display 20 to display the first text information I1 and the fact-check information IFC (step S5). Thereby, the user can visually check the first text Tx1 and perform a fact-check by the user based on the fact-check information IFC.

[0086] [Effect] According to the system 1, the system 1 suitable for generating a text related to travel can be provided. More specifically, the first text Tx1 includes a plurality of items S1 to SN related to travel. Therefore, the information included in the first text Tx1 spans a wide range of fields. Therefore, it is difficult for the LLM to generate the first text Tx1 with a single instruction.

[0087] Therefore, in system 1, the step of the first LLM 318 determining the guidelines for a plurality of second sentences Tx2-1 to Tx2-N, and the step of the second LLM 320 generating the second sentences Tx2-1 to Tx2-N according to the guidelines for the plurality of second sentences Tx2-1 to Tx2-N are executed. Specifically, the control circuit 212 of the first server 210 gives an instruction to the first LLM 318 to generate configuration information IST indicating the guidelines for the plurality of second sentences Tx2-1 to Tx2-N included in each of the plurality of items S1 to SN based on the first definition book information IS1 and the second definition book information IS2. Further, the control circuit 212 of the first server 210 gives an instruction to the second LLM 320 to generate a plurality of second sentence information I2-1 to I2-N indicating the plurality of second sentences Tx2-1 to Tx2-N based on the configuration information IST, the first definition book information IS1, and the second definition book information IS2. Thereby, the first LLM 318 can generate configuration information IST with high accuracy. And the second LLM 320 can generate a plurality of second sentence information I2-1 to I2-N indicating the plurality of second sentences Tx2-1 to Tx2-N based on the configuration information IST with high accuracy. From the above, according to system 1, a system 1 suitable for generating travel-related articles can be provided. Note that in this specification, high accuracy means that the information is highly accurate and / or the content of the information conforms to the user's wishes.

[0088] According to system 1, the first LLM 318 can generate configuration information IST considering the theme of the first sentence Tx1. More specifically, the control circuit 212 of the first server 210 gives an instruction to the first LLM 318 to generate configuration information IST indicating the guidelines for the plurality of second sentences Tx2-1 to Tx2-N included in each of the plurality of items S1 to SN based on the first definition book information IS1, the second definition book information IS2, and the theme information ITE. In this way, the instruction for generating the configuration information IST includes the theme information ITE indicating the theme of the first sentence Tx1. As a result, the first LLM 318 can generate configuration information IST considering the theme of the first sentence Tx1.

[0089] In addition, in System 1, the theme information ITE includes an area input by the user. Therefore, the first LLM 318 can generate configuration information IST considering the area. The area plays a very important role in travel articles. Therefore, the first LLM 318 can generate configuration information IST with high accuracy.

[0090] According to System 1, the first LLM 318 can generate configuration information IST with high accuracy. More specifically, the control circuit 212 of the first server 210 causes the third LLM 322 to convert the project plan information IK into a data exchange format, thereby obtaining the first definition book information IS1 and the second definition book information IS2. In this way, when the first definition book information IS1 and the second definition book information IS2 are described in the data exchange format, the first LLM 318 can accurately understand the contents of the first definition book information IS1 and the second definition book information IS2. As a result, the first LLM 318 can generate configuration information IST with high accuracy based on the first definition book information IS1 and the second definition book information IS2.

[0091] In System 1, the project plan information IK includes information on a plurality of items S1 to SN included in the first text Tx1. The first definition book information IS1 and the second definition book information IS2 are generated based on the project plan information IK. Therefore, the first definition book information IS1 and the second definition book information IS2 include information on a plurality of items S1 to SN. As a result, the first LLM can generate configuration information IST in which information on a plurality of items S1 to SN is reflected based on the first definition book information IS1 and the second definition book information IS2.

[0092] In system 1, the control circuit 212 of the first server 210 gives an instruction to cause the first LLM 318 to generate configuration information IST based on the first definition book information IS1, the second definition book information IS2, the theme information ITE, and the first collection information IR1. The first collection information IR1 is information collected by the fourth LLM based on the first definition book information IS1, the second definition book information IS2, and the theme information ITE. Therefore, the first LLM 318 can reflect the first collection information IR1 collected via the network in the configuration information IST. Also, the first collection information IR1 is information regarding the area and a plurality of spots specified by the theme information ITE. Therefore, the first LLM 318 can reflect the area and the plurality of spots specified by the theme information ITE in the configuration information IST.

[0093] In system 1, the control circuit 212 of the first server 210 uses the sixth LLM 328 to perform fact-checking on a plurality of second text information I2-1 to I2-N. Thereby, the probability of generating a plurality of second text information I2-1 to I2-N including incorrect information is reduced.

[0094] (Other embodiments) Note that the system according to the present invention is not limited to system 1 and can be changed within the scope of its gist.

[0095] Note that the control circuit 212 of the first server 210 (control circuit of one or more computers, second collection information collection means 238) may collect second collection information IR2 using the fifth LLM 326 based on the first definition book information IS1, the second definition book information IS2 (definition book information), the theme information ITE, and the configuration information IST. Then, the control circuit 212 of the first server 210 (control circuit of one or more computers, second text information generation instruction means 224) may give an instruction to cause the second LLM 320 to generate a plurality of second text information I2-1 to I2-N based on the configuration information IST and the second collection information IR2. Thereby, the second LLM 320 can generate a plurality of second text information I2-1 to I2-N based on the second collection information IR2 collected based on the configuration information IST.

[0096] Note that the second collected information IR2 may include external information existing on the Internet and internal information stored in a server with restricted external access. Such internal information may include the first text Tx1 generated in the past. The server with restricted external access is, for example, a data server owned by a travel company. Thereby, the second LLM 320 can generate a plurality of second text information I2-1 to I2-N based on the first text Tx1 generated in the past. Also, as the first text Tx1 generated in the past, an article with a large number of views may be selected.

[0097] Note that when the control circuit 212 (control circuit of one or more computers, second collected information collection means 238) of the first server 210 collects the second collected information IR2 using the fifth LLM 326, it does not have to collect information that does not meet a predetermined condition as the second collected information IR2. Information that does not meet a predetermined condition is, for example, information including an article generated around the time of generating the first text information I1 and having content similar to the first text Tx1. Having content similar to the first text Tx1 means, for example, having the same area.

[0098] Note that when the first text Tx1 is a travel article, at least a part of the plurality of second text information I2-1 to I2-N may include a link to information related to travel products. Travel products are, for example, tours, hotels, and options.

[0099] Note that the second server 310 includes the first LLM 318, the second LLM 320, the third LLM 322, the fourth LLM 324, the fifth LLM 326, and the sixth LLM 328. However, the second server 310 may include the first LLM 318 and may not include the second LLM 320, the third LLM 322, the fourth LLM 324, the fifth LLM 326, and the sixth LLM 328. In this case, the system 1 further includes a fourth server to an eighth server. Each of the fourth server to the eighth server includes the second LLM 320, the third LLM 322, the fourth LLM 324, the fifth LLM 326, and the sixth LLM 328.

[0100] Note that the first LLM 318, the third LLM 322, the fourth LLM 324, the fifth LLM 326, and the sixth LLM 328 may be a single common LLM.

[0101] Note that the flowchart shown in FIG. 6 shows the operations of the control circuit 12 of the first terminal 10 and the control circuit 212 of the first server 210. However, the flowchart of FIG. 6 may be executed, for example, by the control circuit of one computer or by the control circuits of three or more computers.

[0102] Note that the storage circuit 214 of the first server 210 may store the first definition document information IS1 and the second definition document information IS2 in advance. In this case, the project plan information IK is not an essential component. Also, the first definition document information IS1 and the second definition document information IS2 may be generated by the LLM or input by the user.

[0103] Note that the digital document is not limited to a web page. The digital document may be PDF data or WORD data.

[0104] Note that the first definition document information IS1 and the second definition document information IS2 are described in a data exchange format. However, the first definition document information IS1 and the second definition document information IS2 may be text information such as the project plan information IK.

[0105] Note that the first terminal 10 may perform the processes of steps S22 and S23 instead of the first server 210.

[0106] Incidentally, the sixth LLM 328 may evaluate a plurality of second text information I2-1 to I2-N while performing fact-checking on the plurality of second text information I2-1 to I2-N. In this case, a prompt for evaluation is prepared. The prompt for evaluation is input to the sixth LLM 328. The prompt for evaluation determines, for example, whether the plurality of second text information I2-1 to I2-N contains appropriate information. Appropriate information is, for example, whether the plurality of second texts Tx2-1 to Tx2-N are related to the area included in the theme information ITE, or whether they are in line with the content of the plan document information IK, etc. Further, the sixth LLM may evaluate the plurality of second text information I2-1 to I2-N by scoring the plurality of second text information I2-1 to I2-N.

Explanation of Signs

[0107] 1: System 10: First terminal 12: Control circuit 14: Memory circuit 16: Network interface 18: Graphics processing unit 20: Display 26: Operation unit 28: Touch panel 210: First server 212: Control circuit 214: Memory circuit 216: Network interface 220: Definition document information acquisition means 222: Configuration information generation instruction means 224: Second text information generation instruction means 226: Theme information acquisition means 228: Fact-checking means 230: Second text information acquisition means 232: Plan document information acquisition means 234: First collection information collection means 236: First text information generation means 238: Second collection information collection means 310: Second server 312: Control circuit 314: Memory circuit 316: Network interface 320: Second server I1: First article information I2-1~I2-N: Second article information IFC: Fact-checking information II1: First input information II2: Second input information II3: Third input information IK: Proposal information IP: Image information IR1: First collected information IR2: Second collected information IS1: First definition book information IS2: Second definition book information IST: Configuration information ITE: Theme information P1: First prompt P2: Second prompt P3: Third prompt PG: Program Tx1: First article Tx2-1~Tx-N: Second article

Claims

1. The system includes one or more computers. The first sentence is a sentence related to travel and includes a plurality of second sentences; The one or more computer control circuits include: Obtain definition document information indicating a policy of the first document; instructing a first Large Language Model (LLM) to generate configuration information indicating a policy of the plurality of second sentences based on the definition document information; instructing a second LLM to generate a plurality of second sentence information indicating the plurality of second sentences based on the configuration information and the definition document information generated by the first LLM; acquiring the plurality of second document information generated by the second LLM; system.

2. The one or more computer control circuits include: obtaining theme information indicating a theme of the first sentence; instructing a first LLM to generate configuration information indicating a policy of the plurality of second documents based on the definition document information and the theme information; The system of claim 1 .

3. The one or more computer control circuits include: generating first sentence information indicating the first sentence by combining the plurality of second sentences; The system of claim 1 .

4. The one or more computer control circuits include: Acquires the proposal information entered by the user, The business plan information is converted into a data exchange format by a third LLM, thereby acquiring the definition document information. The system of claim 2.

5. The first text is a travel article; The theme information includes an area for input by a user. The system of claim 2.

6. The first text is a travel article; the theme information includes an area for input by a user; The plan information includes a plan for generating a travel article about the area included in the theme information. The system of claim 4.

7. The first sentence includes a plurality of items, The plurality of second sentences correspond to the plurality of items. The system of claim 4.

8. The business plan information includes a plurality of paragraphs, The one or more computer control circuits include: generating, in the third LLM, the definition document information including the plurality of items based on the plurality of paragraphs included in the proposal information; The system of claim 7.

9. The one or more computer control circuits include: Collecting first collected information using a fourth LLM based on the definition book information and the theme information; instructing a first LLM to generate the configuration information based on the definition book information, the theme information, and the first collected information; The system of claim 2.

10. The first text is a travel article; the definition book information includes an instruction to generate a sentence regarding the area specified by the theme information and an instruction to generate a sentence regarding a plurality of spots included in the area specified by the theme information; a control circuit for said one or more computers; Collecting first collected information regarding the area and the plurality of spots identified by the theme information using the fourth LLM; The system of claim 9.

11. The one or more computer control circuits include: Collecting second collected information using a fifth LLM based on the definition book information, the theme information, and the configuration information; instructing the second LLM to generate the plurality of second document information based on the configuration information and the second collected information; The system of claim 2.

12. The second collected information includes external information present on the Internet and internal information stored in a server to which access from outside is restricted, The inside information includes a first sentence generated in the past. The system of claim 11.

13. The one or more computer control circuits include: When collecting second collected information using the fifth LLM, information that does not meet a predetermined condition is not collected as the second collected information. The system of claim 12.

14. The one or more computer control circuits include: using a sixth LLM to fact-check the plurality of second textual information; A system according to any one of claims 1 to 3.

15. The first text is a travel article; At least a portion of the plurality of second text information includes a link to information on a travel product. A system according to any one of claims 1 to 3.

16. the configuration information includes instructions for causing the second LLM to output one or more image information related to the plurality of second documents; At least a portion of the plurality of second text information includes the one or more image information. A system according to any one of claims 1 to 3.

17. the configuration information includes an instruction to cause the second LLM to output URLs referenced in generating the plurality of second sentences; At least a portion of the plurality of second sentence information includes a URL referenced by the second LLM when the plurality of second sentences are generated. A system according to any one of claims 1 to 3.

18. A program, The first sentence is a sentence related to travel and includes a plurality of second sentences; The program includes: obtaining definition document information indicating a policy of the first document; instructing a first Large Language Model (LLM) to generate configuration information indicating a policy of the plurality of second sentences based on the definition document information; instructing a second LLM to generate a plurality of second sentence information indicating the plurality of second sentences based on the configuration information and the definition document information generated by the first LLM; acquiring the plurality of second document information generated by the second LLM; program.

19. A method for producing text information, comprising: The first sentence is a sentence related to travel and includes a plurality of second sentences; The method for producing text information includes: obtaining definition document information indicating a policy of the first document; instructing a first Large Language Model (LLM) to generate configuration information indicating a policy of the plurality of second sentences based on the definition document information; instructing a second LLM to generate a plurality of second sentence information indicating the plurality of second sentences based on the configuration information and the definition document information generated by the first LLM; acquiring the plurality of second document information generated by the second LLM; Text information production method.

Images