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

Abstract

To provide an information processor for supporting the generation of a sentence using a language model.SOLUTION: The above problem is solved by an information processing device including chatbot control means for asking one or more questions to an end user according to a scenario and acquiring an answer to the question, prompt generation means for generating a prompt in which one or more variable portions included in a template prompt are replaced with an answer corresponding to the variable portion, and sentence data acquisition means for acquiring sentence data generated by inputting the generated prompt to a language model, in which the chatbot control means displays the acquired sentence data on a user terminal operated by the end user.SELECTED DRAWING: Figure 3

Description

[Technical Field]

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

[0002] In recent years, various language models have been developed that are capable of not only recognizing and understanding text but also extracting and classifying information from data containing large amounts of text, such as documents, and generating text directly.

[0003] Language models are actively used in various fields, such as search engines, document creation, free dialogue on various topics, data parsing of a given text, providing expert knowledge, programming, and style conversion of a given text, all of which are text-based (see, for example, Patent Document 1). Summary of the Invention [Problem to be solved by the invention]

[0004] Users may create documents (such as event planning documents, product or commodity descriptions, and real estate property introduction documents) for promoting their company's products or services (public relations). Users can easily generate documents by using a language model. However, in order to generate more accurate documents using a language model, it is necessary to appropriately generate prompts to be input into the language model. It can be difficult for users who are not familiar with language models to generate appropriate prompts. Patent Document 1 does not solve this problem.

[0005] An embodiment of the present invention has been made in view of the above-mentioned problems, and has an object to provide an information processing device that supports the generation of sentences using a language model. [Means for solving the problem]

[0006] One embodiment of the present invention is an information processing device comprising a chatbot control means for asking an end user one or more questions according to a scenario and obtaining answers to the questions; a prompt generation means for generating a prompt in which one or more variable parts included in a template prompt are replaced with the answers corresponding to the variable parts; and a sentence data acquisition means for acquiring sentence data generated by inputting the generated prompt into a language model, wherein the chatbot control means displays the acquired sentence data on a user terminal operated by the end user. [Effects of the Invention]

[0007] It can assist in generating sentences using language models. [Brief explanation of the drawings]

[0008] [Figure 1] 1 is a configuration diagram of an example of an information processing system according to an embodiment of the present invention. [Figure 2] FIG. 2 is a diagram illustrating an example of a hardware configuration of a computer according to the present embodiment. [Figure 3] 1 is a functional configuration diagram of an example of an information processing system according to an embodiment of the present invention. [Figure 4] FIG. 10 is a diagram illustrating an example of a chatbot configuration DB. [Figure 5] FIG. 2 is a diagram illustrating an example of a scenario DB. [Figure 6] FIG. 10 is a diagram illustrating an example of a scenario managed in JSON format. [Figure 7] FIG. 10 is a diagram illustrating an example of a scenario managed in JSON format. [Figure 8] FIG. 10 is a diagram illustrating an example of a scenario managed in JSON format. [Figure 9] FIG. 10 is an image diagram of an example of a scenario screen. [Figure 10] FIG. 10 is an image diagram of an example of a chatbot screen. [Figure 11]This is an image of a chat bolt screen showing an example of the content of a conversation between a chatbot function and an end user according to a scenario. [Figure 12] FIG. 10 is a sequence diagram showing an example of a process for generating text data in the information processing system according to the present embodiment. [Figure 13] FIG. 10 is an image diagram of an example of a chatbot screen in which an answer from an end user is selected. [Figure 14] FIG. 1 is an explanatory diagram illustrating an example of an outline of the operation of a chatbot system capable of manned response. DETAILED DESCRIPTION OF THE INVENTION

[0009] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. [First embodiment] <System configuration> 1 is a configuration diagram of an example of an information processing system according to this embodiment. The information processing system 1 according to this embodiment has a configuration in which an information processing device 10 such as a chatbot server, an information processing device 12 such as a text generation server, an administrator terminal 14, and a user terminal 16 are communicatively connected via a network 18. The network 18 is the Internet, a LAN (Local Area Network), or the like.

[0010] The information processing device 10 provides a chatbot function. A chatbot is a term combining chat and bot (robot), and refers to an application or program such as an automatic conversation program that automates certain tasks or processes. Chatting refers to exchanging conversations by inputting text. Note that text may be input using a keyboard or by voice.

[0011] The information processing device 10 may function as a web server and provide data of a web page on which a chatbot is installed to the administrator terminal 14 and the user terminal 16. For example, the information processing device 10 answers questions from the user terminal 16 based on a scenario.

[0012] A scenario is information in which a pre-registered conversation flow is set. The scenario also includes questions (messages) for the end user to respond with text data, as described below. The information processing device 10 performs chatbot processing, which allows the end user to have a conversation with the chatbot function provided by the information processing device 10.

[0013] The information processing device 12 provides a sentence generation function using a language model such as a large language model. The language model used by the information processing device 12 may be, for example, ChatGPT (registered trademark). The information processing device 12 inputs the received prompt into the language model, generates sentence data according to the prompt, and provides the sentence data to the information processing device 10. A prompt is a query or instruction statement that specifies the conditions for the sentence to be generated by the language model. In order to have the language model generate sentence data with high accuracy, it is necessary to devise a prompt to be generated.

[0014] The administrator terminal 14 is operated by an administrator user. The administrator user is, for example, a user of an organization such as a company that provides various services such as chatbot services using the information processing device 10. The administrator terminal 14 is a PC (Personal Computer), a smartphone, a tablet terminal, a mobile phone, a PDA (Personal Digital Assistant), or the like that is operated by the administrator user.

[0015] The user terminal 16 is operated by an end user. The end user is a user who uses various services such as a chatbot service provided by the information processing device 10. The user terminal 16 is a PC, smartphone, tablet terminal, mobile phone, PDA, or the like operated by the end user. The user terminal 16 displays the chatbot function provided by the information processing device 10 and the contents of the conversation between the end user and the user in accordance with the chatbot processing of the information processing device 10.

[0016] In the information processing system 1 according to this embodiment, the information processing device 10 asks an end user one or more questions according to a scenario, acquires answers to the questions, and generates prompts using the acquired answers and template prompts described below. The information processing device 12 generates text data by inputting the prompts generated by the information processing device 10 into a language model, and provides the text data to the information processing device 10. The information processing device 10 then displays the text data provided by the information processing device 12 on the user terminal 16, allowing the end user to confirm it.

[0017] In addition, the administrator terminal 14 or the user terminal 16 may be an output device such as a printer, scanner, facsimile, multifunction peripheral (MFP), projector, display device with electronic whiteboard function, digital signage, etc., a HUD (Head Up Display) device, industrial machinery, imaging device, sound collection device, medical equipment, network home appliance, automobile (Connected Car), game console, etc.

[0018] The information processing devices 10 and 12 are PCs, workstations, etc. The information processing devices 10 and 12 may also be printers, scanners, facsimiles, multifunction peripherals, projectors, display devices with electronic whiteboard functions, output devices such as digital signage, HUD devices, industrial machinery, imaging devices, sound collection devices, medical equipment, network home appliances, automobiles, game consoles, etc.

[0019] The configuration of the information processing system 1 shown in FIG. 1 is one example. The information processing devices 10 and 12 have various system configuration examples depending on the application and purpose. For example, the information processing devices 10 and 12 may be configured as one or more information processing devices. Furthermore, the information processing devices 10 and 12 may be realized as a cloud computing service. The services provided by the information processing devices 10 and 12 may be provided as a single service.

[0020] <Hardware configuration> "computer" The information processing device 10, the information processing device 12, the administrator terminal 14, and the user terminal 16 shown in Fig. 1 are realized by, for example, a computer 500 having a hardware configuration shown in Fig. 2. Fig. 2 is a diagram showing the hardware configuration of an example of a computer according to this embodiment.

[0021] The computer 500 includes a CPU (Central Processing Unit) 501, a ROM (Read Only Memory) 502, a RAM (Random Access Memory) 503, a HD 504, a HDD (Hard Disk Drive) controller 505, a display 506, an external device connection I / F (Interface) 508, a network I / F 509, a data bus 510, a keyboard 511, a pointing device 512, a DVD-RW (Digital Versatile Disk Rewritable) drive 514, and a media I / F 516.

[0022] The CPU 501 controls the overall operation of the computer 500 in accordance with a program. The ROM 502 stores programs used to drive the CPU 501, such as the IPL. The RAM 503 is used as a work area for the CPU 501. The HD 504 stores various data such as programs. The HDD controller 505 controls the reading and writing of various data from and to the HD 504 in accordance with the control of the CPU 501.

[0023] The display 506 displays various types of information such as a cursor, menus, windows, characters, or images. The external device connection I / F 508 is an interface for connecting various types of external devices. In this case, the external device is, for example, a USB (Universal Serial Bus) memory. The network I / F 509 is an interface for data communication using the network 18. The data bus 510 is an address bus, data bus, or the like for electrically connecting each component such as the CPU 501.

[0024] The keyboard 511 is a type of input means having multiple keys for inputting characters, numbers, various instructions, etc. The pointing device 512 is a type of input means for selecting and executing various instructions, selecting a processing target, moving a cursor, etc. The DVD-RW drive 514 controls reading and writing of various data from a DVD-RW 513, which is an example of a removable recording medium. Note that this is not limited to a DVD-RW, and may be a DVD-R, etc. The media I / F 516 controls reading and writing (storing) of data from a recording medium 515, such as a flash memory.

[0025] The hardware configuration shown in FIG. 2 is an example, and it is not necessary to include all of the components shown in FIG. 2, or it may include components other than those shown in FIG.

[0026] <Functional configuration> 3 is a functional configuration diagram of an example of an information processing system according to this embodiment. Note that the functional configuration diagram of FIG. 3 appropriately omits configurations that are not necessary for explaining this embodiment.

[0027] The information processing device 10 realizes the functional configuration of Fig. 3 by, for example, executing an OS (Operating System) and programs on the computer 500 of Fig. 2. The information processing device 10 shown in Fig. 3 includes a chatbot control unit 30, a scenario editing unit 32, a prompt generation unit 34, a text data acquisition unit 36, a chatbot configuration DB 38, and a scenario DB 40.

[0028] 2 executes an OS and a program, the information processing device 12 realizes the functional configuration of Fig. 3. The information processing device 12 shown in Fig. 3 includes a prompt input accepting unit 50, a text data generating unit 52, a text data providing unit 54, and a language model 56.

[0029] 2 executes an OS and a program, the administrator terminal 14 realizes the functional configuration of FIG. 3. The administrator terminal 14 shown in FIG.

[0030] 2 executes an OS and a program, the user terminal 16 realizes the functional configuration of Fig. 3. The user terminal 16 shown in Fig. 3 includes an input unit 90 and a display unit 92.

[0031] The chatbot control unit 30 performs chatbot processing to ask one or more questions to an end user operating the user terminal 16 according to a scenario, and obtain the end user's answers to the questions from the user terminal 16.

[0032] The scenario editing unit 32 displays a scenario screen that accepts scenario editing operations from an administrator user on the administrator terminal 14 operated by the administrator user, and performs processing to edit the scenario in accordance with the administrator user operations accepted on the scenario screen. Details of the scenario editing processing will be described later.

[0033] As will be described later, the prompt generation unit 34 generates a prompt using a template prompt that includes one or more variable parts. The prompt generation unit 34 generates a prompt by replacing one or more variable parts included in the template prompt with an answer that corresponds to the one or more variable parts acquired by the chatbot control unit 30. Details of the processing by the prompt generation unit 34 will be described later.

[0034] The sentence data acquisition unit 36 ​​transmits the prompt generated by the prompt generation unit 34 to the information processing device 12, and acquires sentence data generated (output from the language model 56) by inputting the generated prompt into the language model 56.

[0035] The chatbot configuration DB38 has a configuration shown in Figure 4, for example, and is an example of information for managing chatbot configuration. The item "ID" in the chatbot configuration DB38 is an example of information for identifying a scenario, and may be unique information. The item "ShowText" is an example of text to be displayed on a button by the chatbot. The item "CustomScenarioID" is an example of an ID of a scenario to be displayed on the screen of the administrator terminal 14. Note that "$category" and "$frequent", which are set as examples in the item "CustomScenarioID" in Figure 4, are IDs of built-in scenarios. "17", which is set as an example in the item "CustomScenarioID" in Figure 4, is the ID of a scenario (custom scenario) set on the scenario screen described below.

[0036] The scenario DB 40 has a configuration shown in FIG. 5, for example, and is an example of information for managing scenarios. The item "ID" in the scenario DB 40 is an example of information for identifying a scenario. The item "Flow" manages scenarios set (defined) on a scenario screen, which will be described later. For example, in the column of the item "Flow," scenarios are managed in the Json format shown in FIGS. 6 to 8. FIGS. 6 to 8 are configuration diagrams of an example of scenarios managed in the Json format. Details of the scenarios shown in FIGS. 6 to 8 will be described later.

[0037] The prompt input receiving unit 50 receives a prompt from the information processing device 10 and accepts the input of the prompt. The text data generating unit 52 generates text data according to the prompt by inputting the prompt received from the information processing device 10 into a language model 56. The text data providing unit 54 provides the text data generated by the text data generating unit 52 to the information processing device 10.

[0038] The input unit 70 of the administrator terminal 14 accepts operations from the administrator user. The display unit 72 displays screens such as a scenario screen for the administrator user to confirm. The input unit 90 of the user terminal 16 accepts operations from the end user. The display unit 92 displays the content of the conversation between the chatbot function provided by the information processing device 10 and the end user, and text data such as text generated by the information processing device 12 in response to prompts.

[0039] <Processing> <Scenario Management> The administrator user operates the administrator terminal 14 to display, for example, a scenario screen 1000 shown in FIG. 9, and creates or modifies (hereinafter, referred to as editing) a scenario (hereinafter, referred to as a custom scenario).

[0040] FIG. 9 is an image diagram of an example of a scenario screen. For example, an administrator user logs in to an administration screen or the like and transitions to a scenario screen 1000. The scenario screen 1000 has an edit field 1010 for a custom scenario. The administrator user edits the custom scenario using cards 1020-1 to 1020-7 displayed in the edit field 1010 of FIG. 9, for example. Note that cards 1020-1 to 1020-7 may be collectively referred to as cards 1020. The administrator user can add a new card 1020 to the edit field 1010 by pressing button 1012.

[0041] There are three types of cards 1020: the first card to be executed, the second or subsequent card to be executed, and the card that executes sentence generation using a language model on its backside. Card 1020-1 is an example of the first card to be executed. Cards 1020-2, 1020-3, 1020-5, 1020-6, and 1020-7 are examples of the second or subsequent card to be executed. Card 1020-4 is an example of the card that executes sentence generation using a language model on its backside.

[0042] The scenario screen 1000 also has a template prompt input field 1002. For example, the following template prompt is displayed as an example in the template prompt input field 1002 in Fig. 9:

[0043] "You are an internal public relations officer. Please write a description of the following product in about 150 words that explains its appeal. The product is {{request1}}, the price is {{request2}}, and it has the following features: {{request3}}." 9, the template prompt includes one or more variable parts {{request1}} to {{request3}}. In the edit field 1010, the administrator user edits a scenario for the end user to respond with text data to be substituted for one or more variable parts {{request1}} to {{request3}} of the template prompt.

[0044] For example, card 1020-1 is a card for prompting the end user to respond with text data to be substituted for the variable portion {{request1}} of the template prompt, and for replacing the variable portion {{request1}} of the template prompt.

[0045] Card 1020-2 is a card for prompting the end user to respond with text data to be substituted for the variable portion {{request2}} of the template prompt, and for replacing the variable portion {{request2}} of the template prompt.

[0046] Card 1020-3 is a card for prompting the end user to respond with text data to be substituted for the variable portion {{request3}} of the template prompt, and for replacing the variable portion {{request3}} of the template prompt.

[0047] 9 are replaced with text data of the end user's answer using cards 1020-1 to 1020-3, for example. The sentence data generated by inputting the prompt into the language model 56 is saved in the variable of the output variable input field 1004.

[0048] Card 1020-5 is a card for acquiring end user evaluation information on the sentence data generated by inputting a prompt into the language model 56, and for changing subsequent processing based on the end user evaluation information. If the end user evaluation information on the sentence data generated by inputting a prompt into the language model 56 is "satisfied," processing transitions to returning to the beginning of card 1020-6. If the end user evaluation information on the sentence data generated by inputting a prompt into the language model 56 is not "satisfied," processing transitions to card 1020-7.

[0049] Card 1020-7 is a card for prompting the end user to respond to any complaints the end user has about the sentence data generated by inputting the prompt into the language model 56, and for adding the end user's complaints to the prompt. After processing card 1020-7, processing returns to card 1020-4. The sentence data generated by inputting the corrected prompt into the language model 56 has the complaints answered by the end user improved by processing card 1020-7.

[0050] The administrator user operates the administrator terminal 14 to edit the scenario shown in FIGS. 6 to 8, for example, on a scenario screen 1000 shown in FIG.

[0051] In the JSON format definitions in Figures 6 to 8, the following key-value pairs are saved as an array.

[0052] Key:cardx(uuid):uuid Value: Card type: type scenario-start-card: The first card to be executed scenario-card: Cards to be executed from the second onwards system-card: Executes sentence generation using a language model in the background ports:Action object details id: UUID assigned to each action object type: Type of action object interview-action-object-output-port: Can be used on scenario-start-card or scenario-card Saves the end user input to the variable specified by variableName. text-action-object-output-port: Can be used on scenario-start-card or scenario-card Displays the text specified in text as the button name. backtofirst-action-object-output-port: Press to go to the first option Displays the text specified in text as the button name. gpt-action-object-output-port: Action object to query ChatGPT request: The question to ChatGPT By using {{variable name}}, you can display the value stored in the variable. response: The variable name to save the ChatGPT response next: ID of the next card If empty, end the custom scenario and go to the first option. message: Message to display in the chatbot Can be used on scenario-start-card or scenario-card By using {{variable name}}, you can display the value stored in the variable.

[0053] The administrator user can edit the scenario shown in Figures 6 to 8 by operating the administrator terminal 14, and then select to make the scenario available to end users from the management screen, etc. The end user can select to execute the scenario that has been made available to the user from, for example, the chatbot screen 1200 shown in Figure 10.

[0054] <Scenario execution> 10 is an image diagram of an example of a chatbot screen. By operating the user terminal 16 and pressing a button 1202 on the chatbot screen 1200, the end user can have a conversation with the chatbot function according to a scenario such as those shown in FIGS. 6 to 8.

[0055] Figure 11 is an image diagram of a chat vault screen showing an example of the content of a conversation between a chatbot function and an end user according to a scenario. Messages 1212-1 to 1212-7 on the left side of the chatbot screen 1200 in Figure 11 are examples of messages input by the chatbot function. Messages 1214-1 to 1214-6 on the right side of the chatbot screen 1200 are examples of messages input by an end user.

[0056] Messages 1212-1 and 1214-1 are examples in which the chatbot function asks a question according to the scenario set by card 1020-1, and the end user answers. Based on the conversation content of messages 1212-1 and 1214-1, "potato chips" is saved in variable {{request1}}.

[0057] Messages 1212-2 and 1214-2 are examples in which the chatbot function asks a question according to the scenario set by card 1020-2, and the end user answers. Based on the conversation content of messages 1212-2 and 1214-2, "100 yen" is saved in variable {{request2}}.

[0058] Messages 1212-3 and 1214-3 are examples of the chatbot function asking questions and the end user answering them according to the scenario set by card 1020-3. Based on the conversation content of messages 1212-3 and 1214-3, "Wasabi flavor, super spicy" is saved in variable {{request3}}.

[0059] After message 1214-3 is displayed, the prompt generation unit 34 of the information processing device 10 generates the following prompt in accordance with the contents of the scenario set by card 1020-4, in which the variable parts {{request1}} to {{request3}} included in the template prompt of Figure 9 are replaced with the text data of the end user's answer using cards 1020-1 to 1020-3, for example, ``potato chips,'' ``100 yen,'' and ``wasabi flavor, super spicy.''

[0060] "You are an internal public relations officer. Please write a description of the following product in about 150 characters that explains its appeal. It is a potato chip product, priced at 100 yen, with a wasabi flavor and extremely spicy characteristics." The text data acquisition unit 36 ​​of the information processing device 10 transmits the prompt generated by the prompt generation unit 34 to the information processing device 12 in accordance with the contents of the scenario set by the card 1020-4. The text data acquisition unit 36 ​​acquires text data, for example, text data described in message 1212-4, which is generated by inputting the generated prompt into the language model 56.

[0061] Messages 1212-4 and 1214-4 are examples of displaying sentence data generated by inputting a prompt into the language model 56, and acquiring end-user evaluation information for the generated sentence data in accordance with the contents of the scenario set by card 1020-5. The end-user evaluation information of "No" is acquired based on the conversation contents of messages 1212-4 and 1214-4.

[0062] Messages 1212-5 and 1214-5 are examples of questions asked by the chatbot function according to the content of the scenario set by card 1020-7, and answers given by the end user. From the conversation content of messages 1212-5 and 1214-5, the end user's complaint "Please emphasize the difficulty more." is acquired.

[0063] After message 1214-5 is displayed, the prompt generation unit 34 of the information processing device 10 modifies the prompt to add the end user's complaint, "Please emphasize the difficulty more," in accordance with the contents of the scenario set by card 1020-7.

[0064] The sentence data acquisition unit 36 ​​of the information processing device 10 transmits the prompt generated by the prompt generation unit 34 according to the content of the modified scenario to the information processing device 12. The sentence data acquisition unit 36 ​​acquires sentence data described in the message 1212-6 generated by inputting the generated prompt into the language model 56.

[0065] Messages 1212-6 and 1214-6 are examples of displaying sentence data generated by inputting a prompt into the language model 56, and acquiring end user evaluation information for the generated sentence data in accordance with the contents of the scenario set by card 1020-5. The end user evaluation information "yes" is acquired based on the conversation contents of messages 1212-6 and 1214-6. Message 1212-7 is an example of performing the initial return process in accordance with the contents of the scenario of card 1020-6.

[0066] 12 is a sequence diagram showing an example of a process for generating text data in the information processing system according to this embodiment. In step S10, the chatbot control unit 30 of the information processing device 10 asks a question according to the scenario to an end user operating the user terminal 16. In step S12, the chatbot control unit 30 acquires the end user's answer to the question asked in step S10 from the user terminal 16.

[0067] In step S14, the chatbot control unit 30 refers to the scenario and determines whether there is a next question. If there is a next question, the chatbot control unit 30 returns to the process of step S10 and repeats the processes of steps S10 to S14.

[0068] If it is determined in step S14 that there is no next question, the prompt generation unit 34 of the information processing device 10 generates a prompt in which one or more variable parts included in the template prompt are replaced with the answer acquired in step S12.

[0069] In step S18, the text data acquisition unit 36 ​​of the information processing device 10 transmits the prompt generated in step S16 to the information processing device 12. In step S20, the text data generation unit 52 of the information processing device 12 inputs the prompt received in step S18 into the language model 56, thereby generating text data in accordance with the prompt.

[0070] In step S22, the text data providing unit 54 of the information processing device 12 provides the text data generated in step S20 to the information processing device 10. The text data acquiring unit 36 ​​of the information processing device 10 acquires the text data generated by the information processing device 12 in step S20.

[0071] In step S24, the chatbot control unit 30 of the information processing device 10 causes the text data acquired in step S22 to be displayed on the user terminal 16, for example, as the message 1212-4 shown in FIG.

[0072] In step S26, the input unit 90 of the user terminal 16 acquires evaluation information for the generated text data from the end user. For example, in the case of message 1212-4 in Fig. 11, the evaluation information for the generated text data is an answer as to whether the end user is satisfied or not. If the answer is not satisfied, the evaluation information for the generated text data prompts the end user to answer about their dissatisfaction.

[0073] In step S28, if the end user's evaluation information for the generated text data is satisfactory, the chatbot control unit 30 terminates the processing of the sequence diagram in Figure 12, and if the end user's evaluation information for the generated text data is not satisfactory, it returns to the processing of step S10 and continues processing.

[0074] The process in the sequence diagram of FIG. 12 is an example. For example, if the text data acquired from the information processing device 12 in step S22 contains prohibited words (such as NG words), the prompt generation unit 34 of the information processing device 10 may modify the prompt so that the prohibited words are not used, and transmit the modified prompt to the information processing device 12 to generate text data. Prohibited words may be registered in advance by an administrator user or an end user. Furthermore, prohibited words may be registered for each type of text to be generated. For example, when generating text data for a real estate advertisement, prohibited words may be registered in accordance with the rules for real estate advertisements. Prohibited words may be registered by an administrator user from an administration screen or the like, or by an end user during a conversation with the chatbot function.

[0075] Furthermore, if the response from the end user is made to be a multiple-choice format, it will be easier for the end user to answer the question from the chatbot. Therefore, the information processing system 1 according to this embodiment may make the response from the end user a multiple-choice format, as shown in message 1300 in Fig. 13. Fig. 13 is an image diagram of an example of a chatbot screen in which the response from the end user is made to be a multiple-choice format.

[0076] Message 1300 in FIGS. 13(A) and 13(B) narrows down the keywords of the question in message 1300 from the answers of the end user contained in messages before message 1300 and displays them as recommended words (recommended answers).

[0077] For example, in message 1300 in Figure 13(A), the chatbot narrows down the keywords by type, time, and region from the end user's answers prior to message 1300, "September 2023," "supermarket," and "metropolitan area," and displays "summer festival," "health," "camping," and "fatigue prevention" as recommended words.

[0078] For example, in message 1300 in Figure 13(B), the chatbot narrows down the keywords by type, time, and region from the end user's answers prior to message 1300, "December 2023," "supermarket," and "Tokyo metropolitan area," and displays "Christmas," "health," "year-end party," and "warm and cozy" as recommended words.

[0079] The process of narrowing down the recommended words from the end user's answers may be performed by generating a prompt including the end user's answers and querying ChatGPT for words to recommend as options for the end user's next answer.

[0080] According to the information processing system 1 of this embodiment, an end user can answer questions from a chatbot, thereby generating prompts that can elicit more accurate answers from a language model 56 such as ChatGPT.

[0081] For example, when an end user who is a farmer creates a promotional message, the end user can generate prompts that can elicit more accurate answers from a language model such as ChatGPT by answering questions from a chatbot.56 End users can easily generate appropriate and attractive promotional messages for vegetables, and by posting these promotional messages at shipping destinations (markets, retail stores, e-commerce sites, etc.) or on their own company websites, they can highlight the appeal of their vegetables and expect to increase sales.

[0082] Furthermore, by answering questions from the chatbot, end users, such as real estate agents, can generate prompts that can elicit more accurate answers from language models such as ChatGPT.56 End users can easily generate appropriate and attractive promotional text for vegetables, and by posting this promotional text on a real estate portal site, they can highlight the appeal of the property and hopefully close the deal. [Other embodiments] The information processing system 1 according to this embodiment may be applied to a chatbot system capable of manned response shown in Fig. 14. Fig. 14 is an explanatory diagram illustrating an example of an outline of the operation of a chatbot system capable of manned response.

[0083] A web browser runs on the user terminal 16, and executes a web application provided by the information processing device 10. The web application implements a chatbot that answers questions from end users.

[0084] (1) For example, an end user inputs a question into a user terminal 16 on which a Web application is running.

[0085] (2) The information processing device 10 understands the content of the question from the end user. The information processing device 10 may be an AI (Artificial Intelligence) that has learned the correspondence between questions and answers.

[0086] (3) The information processing device 10 acquires an answer corresponding to the question from the question and answer information storage unit.

[0087] (4) The information processing device 10 transmits the answer to the user terminal 16.

[0088] For example, if an end user selects to execute a scenario, the information processing device 10, as described above, conducts a conversation between the chatbot function and the end user according to the scenario, generates a prompt from the content of the conversation, and sends the prompt to the information processing device 12 to obtain the text data desired by the end user.

[0089] (5) The information processing device 10 stores the conversation between the chatbot function and the end user in a history information storage unit or the like.

[0090] (6) The end user can input an operation requesting human response into the user terminal 16 and shift to human response. For example, as shown in the first embodiment, the end user may shift to human response if satisfactory text data cannot be obtained even after repeatedly generating text data while correcting the prompt. The shift to human response may be performed when the prompt has been corrected a predetermined number of times or more, when it is detected that a word requesting human response is included in the conversation between the chatbot function and the end user, or when the end user performs an operation requesting human response.

[0091] In this way, the chatbot system not only generates text data using the chatbot function shown in the first embodiment and the content of conversations with end users, but also enables support for end users through human response.

[0092] (7) The operator may use his / her knowledge and research to assist in generating prompts so that the end user can obtain satisfactory text data.

[0093] (8) The administrator user can have the administrator terminal 14 communicate with the information processing device 10 and edit the question and answer information in order to enrich the answers to questions, for example.

[0094] Each function of the above-described embodiments can be realized by one or more processing circuits. Here, the term "processing circuit" in this specification includes a processor programmed to perform each function by software, such as a processor implemented by an electronic circuit, as well as devices such as an ASIC (Application Specific Integrated Circuit), a DSP (Digital Signal Processor), an FPGA (Field Programmable Gate Array), or a conventional circuit module designed to perform each function described above.

[0095] The devices described in the examples merely represent one of multiple computing environments for implementing the embodiments disclosed herein. The present invention is not limited to these embodiments, and the components in these embodiments include those that would be easily conceived by a person skilled in the art, those that are substantially identical, and those that are within the scope of what is called equivalents. Furthermore, various omissions, substitutions, modifications, and combinations of the components can be made without departing from the spirit of the present embodiments.

[0096] For example, aspects of the present invention are as follows. <1> A chatbot control means for asking one or more questions to an end user according to a scenario and obtaining answers to the questions; a prompt generating means for generating a prompt in which one or more variable parts included in a template prompt are replaced with the answer corresponding to the variable parts; a sentence data acquisition means for acquiring sentence data generated by inputting the generated prompt into a language model; and The chatbot control means Displaying the acquired text data on a user terminal operated by the end user. An information processing device characterized by: <2> The scenario includes questions that the end user answers and text data that replaces one or more variable parts included in the template prompt. The above characterized by <1> The information processing device described. <3> and a scenario editing means for displaying a scenario screen on an administrator terminal operated by an administrator user, the scenario screen receiving an operation to edit the scenario from the administrator user, and for editing the scenario in accordance with the operation received by the administrator user on the scenario screen. The aforementioned <1> or <2> The information processing device described. <4> the prompt generation means, when the acquired sentence data includes a prohibited word, modifies the prompt so as not to use the prohibited word; The sentence data acquisition means acquires sentence data generated by inputting the corrected prompt into a language model. The above characterized by <1> ~ <3> 10. The information processing device according to claim 9, wherein: <5> the chatbot control means acquires evaluation information of the end user regarding the acquired text data, the prompt generating means modifies the prompt based on the end user's evaluation information; The sentence data acquisition means acquires sentence data generated by inputting the corrected prompt into a language model. The above characterized by <1> ~ <4> 10. The information processing device according to claim 9, wherein: <6> The chatbot control means selects words to be recommended in answers to subsequent questions based on answers to the questions already obtained from the end user, and includes the selected words in the subsequent questions. The above characterized by <1> ~ <5> 10. The information processing device according to claim 9, wherein: <7> The chatbot control means determines a timing for starting to ask one or more questions to the end user according to the scenario based on the end user's operation or the content of the conversation with the chatbot. The aforementioned <1> ~ <6> 10. The information processing device according to claim 9, wherein: <8> An information processing system in which a user terminal operated by an end user and an information processing device that provides a chatbot function communicate with each other via a network, The information processing device includes: A chatbot control means for asking one or more questions to an end user according to a scenario and obtaining answers to the questions; a prompt generating means for generating a prompt in which one or more variable parts included in a template prompt are replaced with the answer corresponding to the variable parts; a sentence data acquisition means for acquiring sentence data generated by inputting the generated prompt into a language model; and The user terminal a display means for displaying the content of the conversation between the chatbot function and the end user, and the text data acquired by the text data acquisition means; have Information processing system. <9> An information processing method performed by an information processing system in which a user terminal operated by an end user and an information processing device that provides a chatbot function communicate via a network, comprising: The information processing device includes: Asking one or more questions to an end user according to a scenario and obtaining answers to said questions; generating a prompt by replacing one or more variable parts included in a template prompt with the answer corresponding to the variable parts; acquiring sentence data generated by inputting the generated prompt into a language model; and The user terminal a step of displaying the conversation between the chatbot function and the end user and the sentence data acquired in the step of acquiring sentence data; have Information processing methods. <10> In the information processing device, a chatbot control procedure for asking one or more questions to an end user according to a scenario and obtaining answers to said questions; a prompt generation step of generating a prompt in which one or more variable parts included in a template prompt are replaced with the answer corresponding to the variable parts; a sentence data acquisition step of acquiring sentence data generated by inputting the generated prompt into a language model; a display step of displaying the acquired text data on a user terminal operated by the end user; A program to execute. [Explanation of symbols]

[0097] 1. Information Processing Systems 10. Information processing equipment 12 Information processing equipment 14 Administrator terminal 16 User terminals 18 Network 30 Chatbot control section 32 Scenario Editorial Department 34 Prompt Generation Unit 36 Text data acquisition section 38 Chatbot Configuration DB 40 Scenario DB 50 Prompt input acceptance section 52 Text data generation unit 54 Text Data Provider 56 language models 70, 90 Input section 72, 92 display section [Prior art documents] [Patent documents]

[0098] [Patent Document 1] Japanese Patent Application Publication No. 2023-43129

Claims

1. a chatbot control means for asking one or more questions to an end user according to a scenario and obtaining answers to the questions; a prompt generating means for generating a prompt in which one or more variable parts included in a template prompt are replaced with the answer corresponding to the variable parts; a sentence data acquisition means for acquiring sentence data generated by inputting the generated prompt into a language model; and The chatbot control means Displaying the acquired text data on a user terminal operated by the end user. An information processing device characterized by:

2. The scenario includes questions that the end user answers and that are to be replaced with text data that replaces one or more variable parts included in the template prompt.

2. The information processing device according to claim 1,

3. and a scenario editing means for displaying a scenario screen on an administrator terminal operated by an administrator user, the scenario screen receiving an operation to edit the scenario from the administrator user, and for editing the scenario in accordance with the operation received by the administrator user on the scenario screen.

3. The information processing device according to claim 1.

4. the prompt generation means, when the acquired sentence data includes a prohibited word, modifies the prompt so as not to use the prohibited word; The sentence data acquisition means acquires sentence data generated by inputting the corrected prompt into a language model.

3. The information processing device according to claim 1, wherein:

5. the chatbot control means acquires evaluation information of the end user regarding the acquired text data, the prompt generating means modifies the prompt based on the end user's evaluation information; The sentence data acquisition means acquires sentence data generated by inputting the corrected prompt into a language model.

3. The information processing device according to claim 1, wherein:

6. The chatbot control means selects words to be recommended in answers to subsequent questions based on answers to the questions already obtained from the end user, and includes the selected words in the subsequent questions.

3. The information processing device according to claim 1, wherein:

7. The chatbot control means determines the timing at which the chatbot control means starts asking the end user one or more questions according to the scenario, based on the end user's operation or the content of the conversation with the chatbot.

3. The information processing device according to claim 1.

8. An information processing system in which a user terminal operated by an end user and an information processing device that provides a chatbot function communicate with each other via a network, The information processing device includes: a chatbot control means for asking one or more questions to an end user according to a scenario and obtaining answers to the questions; a prompt generating means for generating a prompt in which one or more variable parts included in a template prompt are replaced with the answer corresponding to the variable parts; a sentence data acquisition means for acquiring sentence data generated by inputting the generated prompt into a language model; and The user terminal a display means for displaying the content of the conversation between the chatbot function and the end user, and the text data acquired by the text data acquisition means; have Information processing system.

9. An information processing method performed by an information processing system in which a user terminal operated by an end user and an information processing device that provides a chatbot function communicate via a network, comprising: The information processing device includes: Asking one or more questions to an end user according to a scenario and obtaining answers to said questions; generating a prompt in which one or more variable parts included in a template prompt are replaced with the answer corresponding to the variable parts; acquiring sentence data generated by inputting the generated prompt into a language model; and The user terminal a step of displaying the conversation between the chatbot function and the end user and the sentence data acquired in the step of acquiring sentence data; have Information processing methods.

10. In the information processing device, a chatbot control procedure for asking one or more questions to an end user according to a scenario and obtaining answers to the questions; a prompt generation step of generating a prompt in which one or more variable parts included in a template prompt are replaced with the answer corresponding to the variable parts; a sentence data acquisition step of acquiring sentence data generated by inputting the generated prompt into a language model; a display step of displaying the acquired text data on a user terminal operated by the end user; A program to execute.

Images