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

Abstract

To provide an information processing method, a program and an information processing apparatus that remove constraints on the effect of improvement of a chat system using artificial intelligence.SOLUTION: An information processing method of a network service provided using a network causes a server 30 to: acquire at least one of visitor input information that is input by a visitor of the network service through a visitor client device 20 and administrator input information that is input by an administrator of the network service through an administrator client device 10; generate generation information according to the input information by inputting the input information to a trained model; and register the generation information on a knowledge database, the generation information including at least one of a question about the network service and an answer to the question.SELECTED DRAWING: Figure 3

Description

[Technical field]

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

[0002] In recent years, chat systems using artificial intelligence have been attracting attention. By simply inputting natural language into the chat system, a user can receive a response in accordance with the natural language.

[0003] For example, Patent Document 1 discloses a question answering system that uses artificial intelligence and the like to provide a question answering service that can be used jointly by multiple users at low cost. The system separates the functions of a question answering system into those of a question receiving and answering system and a primary answer system, sends questions received by the question receiving and answering system to the primary answer system, creates a primary answer to the question from the primary answer system and returns it to the question receiving and answering system, and creates a final answer based on the received primary answer and sends it to the question terminal.The primary answer system trains the primary answer system on machine learning multiple expected questions and the primary answers so that a single primary answer will be generated for multiple expected questions with common attributes, and the question receiving and answering system which receives the primary answer creates a final answer that includes all answers to the multiple expected questions with common attributes that correspond to the received primary answer. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2019-160286 A Summary of the Invention [Problem to be solved by the invention]

[0005] However, in the chat system of Patent Document 1, the number of expected questions and first answers is reduced in order to reduce the load on machine learning, which limits the improvement effect of the chat system.

[0006] The object of the present invention is to remove the limitations on the improvement effect of chat systems using artificial intelligence. [Means for solving the problem]

[0007] One aspect of the present invention is An information processing method for a network service provided using a network, comprising: The computer Executing a step of acquiring at least one of visitor input information input by a visitor to the network service and administrator input information input by an administrator of the network service; A step of generating generation information according to the input information by inputting the input information into a trained model is performed; the generated information includes at least one of a question about the network service and an answer to the question; performing a step of registering the generation information in a database; It is an information processing method. [Brief description of the drawings]

[0008] [Figure 1] 1 is a block diagram showing a configuration of an information processing system according to an embodiment of the present invention; [Diagram 2] FIG. 2 is a functional block diagram of the information processing system of FIG. 1. [Diagram 3] FIG. 1 is an explanatory diagram of an overview of the present embodiment. [Figure 4] FIG. 4 is a diagram showing a data structure of a visitor database according to the present embodiment. [Diagram 5] FIG. 2 is a diagram illustrating a data structure of a knowledge database according to the present embodiment. [Figure 6] FIG. 2 is a diagram illustrating a data structure of a chat history database according to the present embodiment. [Figure 7] FIG. 4 is a sequence diagram of a knowledge registration process according to the present embodiment. [Figure 8] 8 is a diagram showing an example of a screen displayed in the information processing of FIG. 7. [Figure 9] FIG. 4 is a sequence diagram of a chat process according to the present embodiment. [Figure 10] 10A and 10B are diagrams illustrating an example of a screen displayed in the information processing of FIG. 9. [Figure 11] FIG. 13 is an explanatory diagram of an overview of Modification 1. [Figure 12] FIG. 13 is a sequence diagram of a knowledge registration process according to the first modified example. [Figure 13] 13 is a diagram showing an example of a screen displayed in the information processing of FIG. 12. [Figure 14] FIG. 11 is an explanatory diagram of an overview of Modification 2. [Figure 15] FIG. 13 is a diagram showing a data structure of a knowledge database in the second modification example. [Figure 16] FIG. 11 is a sequence diagram of a knowledge registration process according to the second modified example. [Figure 17] 17A and 17B are diagrams showing examples of screens displayed in the information processing of FIG. 16. [Figure 18] FIG. 13 is an explanatory diagram of an overview of Modification 3. [Figure 19] FIG. 13 is a sequence diagram of a knowledge registration process according to the third modified example. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings. In the drawings for explaining the embodiment, the same components are generally designated by the same reference numerals, and the repeated description will be omitted.

[0010] (1) Information Processing System Configuration The configuration of the information processing system will be described below. Fig. 1 is a block diagram showing the configuration of the information processing system of this embodiment. Fig. 2 is a functional block diagram of the information processing system of Fig. 1.

[0011] As shown in FIG. 1, the information processing system 1 includes an administrator client device 10, a visitor client device 20, and a server 30. The administrator client device 10 and the server 30 are connected via a network (eg, the Internet or an intranet) NW.

[0012] The administrator client device 10 is a computer (an example of an "information processing device") that transmits a request to the server 30. The administrator client device 10 is, for example, a smartphone, a tablet terminal, or a personal computer.

[0013] The visitor client device 20 is a computer that sends a request to the server 30. The visitor client device 20 is, for example, a smartphone, a tablet terminal, or a personal computer.

[0014] The server 30 is a computer (an example of an "information processing device") that provides the administrator client device 10 or the visitor client device 20 with a response in response to a request transmitted from the administrator client device 10 or the visitor client device 20. The server 30 is, for example, a web server.

[0015] (1-1) Configuration of the administrator client device The configuration of the administrator client device 10 will be described.

[0016] As shown in FIG. 2, the administrator client device 10 includes a storage device 11, a processor 12, an input / output interface 13, and a communication interface 14.

[0017] The storage device 11 is configured to store programs and data. The storage device 11 is, for example, a combination of a Read Only Memory (ROM), a Random Access Memory (RAM), and a storage (for example, a flash memory or a hard disk).

[0018] The programs include, for example, the following programs: ·OS (Operating System) programs Parameters, functions, and models used in information processing · Programs for applications that process information (e.g. web browsers)

[0019] The data includes, for example, the following data: Databases referenced in information processing Data obtained by performing information processing (i.e., the results of performing information processing)

[0020] The processor 12 is configured to realize the functions of the administrator client device 10 by starting a program stored in the storage device 11. The processor 12 is, for example, a central processing unit (CPU), a graphics processing unit (GPU), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or a combination of these.

[0021] The input / output interface 13 is configured to obtain user instructions from an input device connected to the administrator client device 10 , and to output information to an output device connected to the administrator client device 10 . The input device is, for example, a keyboard, a pointing device, a touch panel, or a combination thereof. The output device is, for example, a display.

[0022] The communication interface 14 is configured to control communications between the administrator client device 10 and the server 30 .

[0023] (1-2) Configuration of visitor client device The configuration of the visitor client device 20 will now be described.

[0024] As shown in FIG. 2, the visitor client device 20 includes a storage device 21, a processor 22, an input / output interface 23, and a communication interface 24.

[0025] The storage device 21 is configured to store programs and data, and is, for example, a combination of a ROM, a RAM, and a storage (for example, a flash memory or a hard disk).

[0026] The programs include, for example, the following programs: -OS programs Parameters, functions, and models used in information processing · Programs for applications that process information (e.g. web browsers)

[0027] The data includes, for example, the following data: Databases referenced in information processing Data obtained by performing information processing (i.e., the results of performing information processing)

[0028] The processor 22 is configured to realize the functions of the visitor client device 20 by running a program stored in the storage device 21. The processor 22 is, for example, a CPU, a GPU, an ASIC, an FPGA, or a combination thereof.

[0029] The input / output interface 23 is configured to obtain user instructions from input devices connected to the visitor client device 20 and to output information to output devices connected to the visitor client device 20 . The input device is, for example, a keyboard, a pointing device, a touch panel, or a combination thereof. The output device is, for example, a display.

[0030] The communications interface 24 is configured to control communications between the visitor client device 20 and the server 30 .

[0031] (1-3) Server configuration The configuration of the server 30 will now be described.

[0032] As shown in FIG. 2, the server 30 includes a storage device 31, a processor 32, an input / output interface 33, and a communication interface .

[0033] The storage device 31 is configured to store programs and data, and is, for example, a combination of a ROM, a RAM, and a storage (for example, a flash memory or a hard disk).

[0034] The programs include, for example, the following programs: -OS programs Parameters, functions, and models used in information processing Application programs that perform information processing

[0035] The data includes, for example, the following data: Databases referenced in information processing Results of information processing

[0036] The processor 32 is configured to realize the functions of the server 30 by running a program stored in the storage device 31. The processor 32 is, for example, a CPU, a GPU, an ASIC, an FPGA, or a combination thereof.

[0037] The input / output interface 33 is configured to receive user instructions from an input device connected to the server 30 , and to output information to an output device connected to the server 30 . The input device is, for example, a keyboard, a pointing device, a touch panel, or a combination thereof. The output device is, for example, a display.

[0038] The communication interface 34 is configured to control communications between the server 30 and the administrator client device 10 .

[0039] (2) Overview of the embodiment An overview of this embodiment will be described below with reference to Fig. 3, which is an explanatory diagram of the overview of this embodiment.

[0040] 3, the administrator provides administrator input information to the server 30 via the administrator client device 10. The administrator is, for example, a person in charge of a company that provides commercial materials (at least one of goods and services) to the market and operates a network service. A visitor provides visitor input information to the server 30 via the visitor client device 20. A visitor is, for example, a person who asks a question about a product provided by the administrator via the network service.

[0041] The server 30 generates generation information (questions and answers) using at least one of the administrator input information and the visitor input information. The server 30 registers the generated information in a knowledge database.

[0042] (3) Database The databases of this embodiment are stored in the storage device 31.

[0043] (3-1) Visitor database The visitor database of this embodiment will now be described with reference to Fig. 4, which is a diagram showing the data structure of the visitor database of this embodiment.

[0044] Visitor information is stored in the visitor database in Fig. 4. The visitor information is information about visitors to the network service. The visitor database includes a "visitor ID" field, a "visitor name" field, and a "visitor attribute" field. Each field is associated with the other fields.

[0045] The "visitor ID" field stores visitor identification information. The visitor identification information is information that identifies a visitor.

[0046] The "visitor name" field stores visitor name information. The visitor name information is information about the visitor's name.

[0047] The "visitor attribute" field stores visitor attribute information. The visitor attribute information is information about the attributes of the visitor.

[0048] The "visitor attribute" field includes a "gender" field and an "age" field.

[0049] The "gender" field stores gender information. The gender information is information about the gender of the visitor.

[0050] The "age" field stores age information. The age information is information about the age of the visitor.

[0051] (3-2) Knowledge Database The knowledge database of this embodiment will now be described with reference to Fig. 5, which shows the data structure of the knowledge database of this embodiment.

[0052] Knowledge information is stored in the knowledge database of Fig. 5. The knowledge information is a combination of questions and answers to inquiries about network services. The knowledge database includes a "knowledge ID" field, a "question" field, an "answer" field, a "knowledge category" field, and a "tag" field.

[0053] The "knowledge ID" field stores knowledge identification information. The knowledge identification information is information for identifying the knowledge information.

[0054] The "Question" field stores question information. The question information is information related to the question of the inquiry.

[0055] The "answer" field stores answer information. The answer information is information about an answer to a question of an inquiry.

[0056] The "knowledge category" field stores knowledge category information. The knowledge category information is information related to a knowledge category. A knowledge category is a type, genre, or classification of knowledge information. The knowledge category information includes, for example, at least one of the following: ·application ·Fee ·function Cancellation

[0057] The "tag" field stores tag information. Tag information is information about tags for combinations of questions and answers. Tag information is information used as a search target. Tag information is set arbitrarily by an administrator, or is set using a predetermined tag model (a model that generates at least one synonym for a question and an answer). The tag information is, for example, information relating to at least the following: Synonyms of the question keywords · Synonyms of answer keywords

[0058] (3-3) Chat history database The chat history database of this embodiment will be described with reference to Fig. 6, which is a diagram showing the data structure of the chat history database of this embodiment.

[0059] The chat history database in FIG. 6 stores chat history information. The chat history information is information about the history of chats. The chat history database includes a "thread ID" field, a "timestamp" field, a "post" field, a "poster" field, a "chat category" field, and a "rating" field. One record in the chat history database corresponds to a collection of posts (i.e., a thread). The chat history database is associated with the visitor identification information.

[0060] The "thread ID" field stores thread identification information. The thread identification information is information for identifying a thread. A thread is a collection of multiple posts. A thread is a set of posts made from when a visitor gives an instruction to start a chat until when the visitor gives an instruction to end the chat.

[0061] The "timestamp" field stores the post's timestamp (i.e., information about when the post was made). If a thread contains multiple posts, the "timestamp" field stores multiple timestamps, each associated with a different post.

[0062] The "Post" field stores post information. The post information is information about the content of a post made by a poster. The posted information is, for example, information in at least one of the following formats. ·text Still images ·movie ·audio ·icon

[0063] The "Contributor" field stores poster information. The poster information is, for example, information on at least one of the following: If a thread contains multiple posts, the "Contributor" field stores multiple pieces of poster information. Each piece of poster information is associated with each of the multiple posts. Visitor Identification Information - Information that identifies the administrator (hereinafter referred to as "Administrator Identification Information") Information indicating that the response is an automated response from the server 30 (i.e., a chatbot)

[0064] The "chat category" field stores chat category information. If a thread includes multiple posts, the "chat category" field stores multiple pieces of chat category information. Each piece of chat category information is associated with each of the multiple posts.

[0065] The "rating" field stores rating information, which is information about a visitor's rating of a single thread.

[0066] (4) Information processing The information processing of this embodiment will be described.

[0067] (4-1) Knowledge registration process The knowledge registration process of this embodiment will be described below with reference to Fig. 7. Fig. 8 is a diagram showing an example of a screen displayed in the information processing of Fig. 7.

[0068] The trigger in FIG. 7 is a predetermined instruction by the administrator (for example, an instruction to log in using administrator identification information and an operation for displaying screen P1110).

[0069] As shown in FIG. 7, the administrator client device 10 executes a generation request (S1110). The processor 12 displays screen P1110 (FIG. 8) on the display.

[0070] The screen P1110 includes operation objects B1110a to B1110b and field objects F1110a to F1110d. The field objects F1110a to F1110d are objects that accept user instructions to specify input information.

[0071] The field object F1110a is an object that accepts a user instruction for inputting a file (an example of "input information").

[0072] The field object F1110b is an object that receives a user instruction for inputting chat history information (an example of "input information") stored in the chat history database (FIG. 6).

[0073] The field object F1110c is an object that accepts a user instruction for inputting the results of scraping a website or intranet site (an example of "input information").

[0074] The field object F1110d is an object that accepts a user instruction for inputting text (an example of "input information").

[0075] The operation object B 1110a is an object that receives a user instruction to mask information of a type designated by the user (hereinafter, referred to as a "mask type") among input information. The mask type includes, for example, at least one of the following. - Information that can be determined whether it should be masked (hereinafter referred to as "masked information") according to rules established by the administrator (for example, personal information of visitors, personal information of the administrator, confidential information of the administrator) - Information that can be determined whether or not it is maskable according to rules set by the visitor (e.g., the visitor's confidential information)

[0076] The operation object B 1110b is an object that receives a user instruction for generating a summary of input information.

[0077] As a first example, when the administrator inputs input information into at least one of the field objects F1110a to F1110d, the processor 12 transmits generation request data to the server 30. The generation request data includes, for example, the following information: Input information entered into field objects F1110a to F1110d

[0078] As a second example, when the administrator inputs input information into at least one of the field objects F1110a to F1110d and operates the operation object B1110a, the processor 12 transmits generation request data to the server 30. The generation request data includes, for example, the following information: Input information entered into field objects F1110a to F1110d Information regarding the mask type specified in the operation object B 1110a

[0079] After step S1110, the server 30 executes generation (S1130). Specifically, the storage device 31 stores a generative model. A generative model is a trained model trained by machine learning. A generative model is configured to output generated information in response to input information.

[0080] In a first example of step S1130, the processor 32 provides the input information included in the generation request data to the generation model. The generative model outputs generated information in response to input information. The generated information is a question and an answer.

[0081] In a second example of step S1130, two types of mask models (a first mask model and a second mask model) are stored in the storage device 31. The mask models are, for example, learned models learned by machine learning or rule-based models. The mask models are configured to remove mask target information from input information. The first mask model is generated using rules defined by the administrator and data prepared by the administrator. The second mask model is generated using rules defined by the visitor and data provided by the administrator.

[0082] The processor 32 selects a mask model (at least one of the first mask model and the second mask model) corresponding to the information on the mask type included in the generation request data. Processor 32 applies the input information contained in the generation request data to the selected mask model. The mask model outputs information in which the mask target information has been removed from the input information (hereinafter referred to as "masked input information"). The processor 32 provides the mask input information to the generative model. The generative model outputs generated information in response to the masked input information. The generated information is a question and an answer.

[0083] After step S1130, the server 30 executes the generated response (S1131). Specifically, the processor 32 transmits the generated response data to the administrator client device 10. The generated response data includes, for example, the following information: The generated information obtained in step S1130

[0084] After step S1131, the administrator client device 10 executes presentation of the generated result (S1111). Specifically, the processor 12 displays screen P1111 (FIG. 8) on the display. The screen P1111 includes display objects A1111a to A1111b, an operation object B1111, and field objects F1111a to F1111b.

[0085] The display object A 1111a is an object that displays a question from among the generation information included in the generated response data.

[0086] The display object A 1111b is an object that displays a response from among the generated information included in the generated response data.

[0087] The operation object B 1111 is an object that receives a user instruction for registering the generated information included in the generated response data in the knowledge database.

[0088] The field object F1111a is an object that receives input of tag information for generated information.

[0089] The field object F1111b is an object that receives input of knowledge category information for the generated information.

[0090] After step S1111, the administrator client device 10 executes a registration request (S1112). Specifically, when the administrator inputs any tag information into the field object F1111a, inputs any knowledge category information into the field object F1111b, and operates the operation object B1111, the processor 12 transmits registration request data to the server 30. The registration request data includes, for example, the following information: Tag information entered in field object F1111a Knowledge category information entered in field object F1111b

[0091] After step S1112, the server 30 executes updating of the database (S1132). Specifically, the processor 32 adds a new record to the knowledge database (FIG. 5) in response to the registration request data. The following information is stored in each field of the new record: "Knowledge ID" field: New knowledge identification information "Question" field: the question obtained in step S1130 "Answer" field: Answer obtained in step S1130 "Tag" field: tag information included in the registration request data "Knowledge Category" field: Knowledge category information included in the registration request data

[0092] After step S1132, the server 30 executes a registration response (S1133). Specifically, the processor 32 transmits the registration response data to the administrator client device 10 .

[0093] After step S1133, the administrator client device 10 executes a knowledge use request (S1113). Specifically, the processor 12 displays screen P1112 (FIG. 8) on the display. The screen P1112 includes operation objects B1112a to B1112c.

[0094] The operation object B 1112a is an object that receives a user instruction for fine-tuning an existing model using knowledge information.

[0095] The operation object B 1112b is an object that receives a user instruction for generating a new model using knowledge information.

[0096] The operation object B 1112c is an object that receives a user instruction for generating FAQs (Frequently Asked Questions) using knowledge information.

[0097] When the administrator operates any of the operation objects B1112a to B1112c, the processor 12 transmits knowledge utilization request data to the server 30. The knowledge utilization request data includes, for example, the following information: Information indicating an instruction corresponding to any one of the operation objects B1112a to B1112c

[0098] After step S1113, the server 30 executes the use of knowledge (S1134).

[0099] In a first example of step S1134, if the knowledge usage request data includes an instruction corresponding to operation object B1112a, the processor 32 uses knowledge information in the knowledge database (Figure 5) to fine-tune an existing learned model (e.g., a published language model or a language model pre-stored in the storage device 31).

[0100] In the second example of step S1134, if the knowledge use request data includes an instruction corresponding to the operation object B 1112b, the processor 32 uses the knowledge information in the knowledge database (FIG. 5) as learning data to generate a new trained model.

[0101] The trained model obtained in the first or second example of step S1134 can be used as an answer model. The answer model is configured to output an answer according to an input question.

[0102] In a third example of step S1134, an FAQ model is stored in the storage device 31. The FAQ model is configured to generate FAQs corresponding to knowledge information. An FAQ is a set of combinations of questions and answers. In an FAQ, questions and answers have a one-to-one correspondence. Processor 32 inputs questions and answers stored in a knowledge database (FIG. 5) into the FAQ model. The FAQ model generates FAQs (a set of new questions and new answers) that correspond to the input questions and answers. The processor 32 stores the generated FAQ in an FAQ database (not shown).

[0103] After step S1134, the server 30 executes a knowledge use response (S1135). Specifically, the processor 32 transmits the knowledge use response data to the administrator client device 10 .

[0104] After step S1135, the administrator client device 10 executes a completion notification (S1114). Specifically, the processor 12 displays screen P1113 (FIG. 8) on the display. A message indicating that knowledge usage has been completed is displayed on the screen P1113.

[0105] (4-2) Chat processing The chat process of this embodiment will be described with reference to Fig. 9, which is a sequence diagram of the chat process of this embodiment. Fig. 10 is a diagram showing an example of a screen displayed in the information process of Fig. 9.

[0106] The trigger in FIG. 9 is a predetermined instruction by the visitor (for example, an instruction to log in using the visitor identification information and an operation for displaying the screen P1220).

[0107] As shown in FIG. 9, the visitor client device 20 executes the answer request (S1220). Specifically, the processor 22 displays a screen P1220 (FIG. 10) on the display. The screen P1220 includes a field object F1220. A chat input object is displayed in the field object F1220. The chat input object corresponds to, for example, a question frequently asked by visitors to the website.

[0108] When the visitor designates a desired question from among the chat input objects displayed in the field object F1220, the processor 22 transmits answer request data to the server 30. The answer request data includes, for example, the following information: Questions specified by the visitor Visitor Identification Information

[0109] After step S1220, the server 30 executes answer generation (S1230).

[0110] In a first example of step S1230, the storage device 31 stores an answer model. The processor 32 inputs the question contained in the answer request data into the answer model. The answer model outputs an answer according to the input question.

[0111] In a second example of step S1230, the processor 32 transmits a question to the administrator client device 10. The processor 12 presents a question sent from the server 30 . The processor 12 accepts input of a response to the question from the administrator. The processor 12 sends the response to the server 30.

[0112] After step S1230, the server 30 executes an answer generation response (S1231). Specifically, the processor 32 transmits the answer generation response data to the visitor client device 20. The answer generation response data includes, for example, the following information: Answer obtained in step S1230

[0113] After step S1231, the visitor client device 20 executes presentation of an answer (S1221). Specifically, the processor 22 displays the screen P1221 on the display. The screen P1221 includes a display object A1221, an operation object B1221, and a field object F1221.

[0114] The display object A 1221 is an object that displays the answer included in the answer generation response data.

[0115] The operation object B 1221 is an object that receives a user instruction to end the chat.

[0116] The field object F1221 is an object that receives input of posted information.

[0117] When a visitor inputs posting information into the field object F1221, the processor 22 transmits answer request data to the server 30. The answer request data includes the posting information input into the field object F1221.

[0118] The server 30 uses the posted information included in the answer request data to execute answer generation (S1230) to answer generation response (S1231).

[0119] The visitor client device 20 presents an answer (S1221). Specifically, the processor 22 displays screen P1222 (FIG. 8) on the display. Like the screen P1221, the screen P1222 includes a display object A1221, an operation object B1221, and a field object F1221.

[0120] When the visitor operates the operation object B1221, the processor 22 displays a screen P1223 (FIG. 8) on the display. The screen P1223 includes an operation object B1223 and a field object F1223.

[0121] The operation object B 1223 is an object that receives a user instruction for finalizing the end of the chat.

[0122] The field object F1223 is an object that receives an input of an evaluation of the chat by the visitor (for example, "GOOD" or "BAD").

[0123] After step S1221, the visitor client device 20 executes the end request (S1222). Specifically, when the visitor operates the operation object B 1223, the processor 22 transmits the end request data to the server 30. The termination request data includes, for example, the following information: Posted information (i.e., questions entered by the visitor between the start of the chat and the operation of the operation object B1223 (for example, information entered into the field objects F1220 to F1221), and answers entered between the start of the chat and the operation of the operation object B1223 (for example, at least one of the answers generated by the answer model and the answers entered by the administrator) - Rating entered in field object F1223

[0124] After step S1222, the server 30 executes updating of the database (S1232). Specifically, the processor 32 updates the chat history database (FIG. 6). The following information is stored in each field of the chat history database: "Thread ID" field: New thread identification information "Timestamp" field: Information about the date and time the chat was conducted "Poster" field: Poster information (if the chat content is a question, this is the "visitor identification information", and if the chat content is an answer, this is the "chatbot" or the administrator identification information of the administrator who entered the answer) "Post" field: Questions and answers included in the end request data

[0125] (5) Summary of the present embodiment According to this embodiment, generated information (at least one of a question and an answer) is generated using input information (at least one of administrator input information and visitor input information). The generated information is registered in a knowledge database. This allows the administrator to easily increase the amount of information in the knowledge database. In particular, when generating generated information (at least one of questions and answers) using both administrator input information and visitor input information, the administrator can further easily increase the amount of information in the knowledge database. Furthermore, when generating generated information (both questions and answers) using at least one of administrator input information and visitor input information, the administrator can further easily increase the amount of information in the knowledge database. Furthermore, when generating generated information (both questions and answers) using both administrator input information and visitor input information, the administrator can further easily increase the amount of information in the knowledge database. As a result, it is possible to remove limitations on the effectiveness of improvements to chat systems using artificial intelligence.

[0126] According to this embodiment, one piece of generation information (at least one of one question and one answer) may be generated using one piece of input information (at least one of administrator input information and visitor input information). This allows the administrator to more easily increase the amount of information in the knowledge database.

[0127] According to this embodiment, a single piece of input information (at least one of administrator input information and visitor input information) may be used to generate a plurality of pieces of generated information (either a plurality of questions or a plurality of answers). This allows the administrator to more easily increase the amount of information in the knowledge database.

[0128] According to this embodiment, the generated information (at least one of a question and an answer) may be presented to an administrator. In this case, the administrator, after viewing the presented generated information, issues an instruction to register only the desired generated information in the knowledge database. Only the generated information according to the administrator's instruction (i.e., the generated information that the administrator determines should be registered) is registered in the knowledge database. This makes it possible to prevent low-quality generated information (that is, generated information that is not useful for chat processing) from being mixed into the knowledge database.

[0129] According to the present embodiment, knowledge information from a knowledge database may be used to fine-tune the answer model. This enables the administrator to provide the visitor with chat processing using an answer model that is more suitable for the knowledge information.

[0130] According to this embodiment, the knowledge information in the knowledge database may be used as a training data set to generate an answer model. This enables the administrator to provide the visitor with chat processing using an answer model that is more suitable for the knowledge information.

[0131] According to this embodiment, the input information may be a chat history. This makes it possible to obtain generated information that reflects the questions and answers of visitors in the chat.

[0132] According to this embodiment, the generation information may be generated taking into consideration the evaluation of the chat history. This makes it easier to obtain generated information that reflects the evaluation of the chat history (for example, the higher the evaluation, the stronger the influence of the chat history is reflected in the generated information).

[0133] According to this embodiment, the mask target information may be removed from the input information. This makes it possible to prevent information that is inconvenient for the visitor from being transmitted to the server 30 and from being reflected in the generated information, thereby enabling the visitor to feel at ease while chatting.

[0134] (6) Variations A modification of this embodiment will now be described.

[0135] (6-1) Variation 1 A description will be given of Modification 1. Modification 1 is an example in which the generated information is a similar question.

[0136] (6-1-1) Overview of Modification 1 An overview of Modification 1 will be described below. Fig. 11 is an explanatory diagram of an overview of Modification 1.

[0137] As shown in FIG. 11, the administrator provides administrator input information to the server 30 via the administrator client device 10 . A visitor provides visitor input information to server 30 via visitor client device 20 .

[0138] The server 30 generates generated information (for example, at least one of a question type and an answer type) using at least one of the administrator input information and the visitor input information. The server 30 registers the generated information in a knowledge database.

[0139] (6-1-2) Knowledge registration process of modification example 1 The following describes the knowledge registration process of Modification 1. Fig. 12 is a sequence diagram of the knowledge registration process of Modification 1. Fig. 13 is a diagram showing an example of a screen displayed in the information processing of Fig. 12.

[0140] The trigger in FIG. 12 is a predetermined instruction by the administrator (for example, an operation for displaying screen P2110).

[0141] As shown in FIG. 12, the administrator client device 10 executes a generation request (S2110). The processor 12 displays screen P2110 (FIG. 13) on the display.

[0142] The screen P2110 includes field objects F2110a to F2110b. The field objects F2110a to F2110b are objects that accept user instructions that specify input information.

[0143] The field object F2110a is an object that receives a user instruction for inputting a free word.

[0144] The field object F2110b is an object that receives a user instruction for inputting a class name of a free word (an example of "input information").

[0145] When the administrator inputs input information into the field objects F2110a to F2110b, the processor 12 transmits generation request data to the server 30. The generation request data includes, for example, the following information: Input information entered into field objects F2110a to F2110b

[0146] After step S2110, the server 30 executes generation (S2130). Specifically, the storage device 31 stores a generative model. The generative model is a trained model trained by machine learning and configured to output similar questions corresponding to free words.

[0147] The processor 32 provides the free words included in the generation request data to the generation model. The generative model outputs a similar question corresponding to the free word. A similar question is a string of characters that can be linguistically interpreted in the same way as the free word by the generative model.

[0148] After step S2130, the server 30 executes the generated response (S2131). Specifically, the processor 32 transmits the generated response data to the administrator client device 10. The generated response data includes, for example, the following information: Similar questions obtained in step S2130

[0149] After step S2131, the administrator client device 10 executes presentation of the generation result (S2111). Specifically, the processor 12 displays screen P2111 (FIG. 13) on the display.

[0150] The screen P2111 includes a display object A2111 and an operation object B1111.

[0151] Display object A 2111 is an object that displays a similar question included in the generated response data.

[0152] The operation object B 2111 is an object that receives a user instruction for registering a similar question included in the generated response data in the knowledge database.

[0153] After step S2111, the administrator client device 10 executes a registration request (S2112). Specifically, when the administrator operates the operation object B 2111, the processor 12 transmits registration request data to the server 30.

[0154] After step S2112, the server 30 executes updating of the database (S2132). Specifically, the processor 32 adds a new record to the knowledge database (FIG. 5) in response to the registration request data. The following information is stored in each field of the new record: "Knowledge ID" field: New knowledge identification information "Question" field: Similar questions obtained in step S2130 "Answer" field: Similar questions obtained in step S2130

[0155] (6-1-3) Summary of Modification 1

[0156] According to the first modification, similar questions may be generated using input information. This allows the administrator to increase the amount of information in the knowledge database without increasing the number of steps required to input information to the knowledge database, thereby further removing limitations on the effect of improving chat systems using artificial intelligence.

[0157] According to the first modification, the generated information (similar questions) may be presented to an administrator. In this case, the administrator, after viewing the presented similar questions, instructs the system to register only the desired similar questions in the knowledge database. Only similar questions that comply with the administrator's instructions (i.e., similar questions that the administrator determines should be registered) are registered in the knowledge database. This makes it possible to prevent low-quality generated information (i.e., topics that are not useful for chat processing) from being mixed into the knowledge database.

[0158] (6-2) Variation 2 A description will now be given of Modification 2. Modification 2 is an example in which the generated information is a summary of the chat.

[0159] (6-2-1) Overview of Modification 2 An overview of Modification 2 will be described below. FIG 14 is an explanatory diagram of an overview of Modification 2.

[0160] As shown in FIG. 14, the administrator provides chat history information to the server 30 via the administrator client device 10 .

[0161] The server 30 generates chat summary information using the chat history information. The chat summary information is information about a summary of a chat based on the chat history information. The server 30 registers the chat summary information in a knowledge database.

[0162] (6-2-2) Knowledge database of modification example 2 A description will now be given of the knowledge database of the modified example 2. Fig. 15 is a diagram showing the data structure of the knowledge database of the modified example 2.

[0163] Knowledge information is stored in the knowledge database of Fig. 15. The knowledge information is chat summary information of chat inquiries about network services. The knowledge database includes a "knowledge ID" field, a "summary" field, a "tag" field, and a "class name" field. The “Knowledge ID” field, the “Tag” field, and the “Class Name” field are the same as those in FIG.

[0164] The "Summary" field stores chat summary information.

[0165] (6-2-3) Knowledge registration process of modification example 2 A description will be given of chat processing according to Modification 2. Fig. 16 is a sequence diagram of the knowledge registration processing according to Modification 2. Fig. 17 is a diagram showing an example of a screen displayed in the information processing of Fig. 16.

[0166] The trigger in FIG. 16 is similar to that in FIG.

[0167] As shown in FIG. 16, the administrator client device 10 executes a generation request (S3110). Specifically, the processor 12 displays a screen P1110 (FIG. 8) on the display. When the administrator inputs a chat history into the field object F 1110b and operates the operation object B 1110b, the processor 12 transmits generation request data to the server 30. The generation request data includes, for example, the following information: Chat history information entered in field object F1110b

[0168] After step S3110, the server 30 executes generation (S3130). Specifically, the storage device 31 stores a generative model. The generative model is a trained model trained by machine learning. The generative model is configured to output chat summary information (an example of “generated information”) according to chat history information.

[0169] In a first example of step S3130, the processor 32 provides the chat history information included in the generation request data to the generation model. The generative model outputs generated information corresponding to the chat history information, the generated information being chat summary information.

[0170] In the second example of step S3130, the storage device 31 stores two types of mask models (a first mask model and a second mask model).

[0171] The processor 32 selects a mask model (at least one of the first mask model and the second mask model) corresponding to the information on the mask type included in the generation request data. Processor 32 applies the chat history information included in the generation request data to the selected mask model. The mask model outputs information in which mask target information has been removed from chat history information (hereinafter referred to as "mask input information"). The processor 32 provides the mask input information to the generative model. The generative model outputs generated information corresponding to the masked input information, the generated information being chat summary information.

[0172] The generated response (S1131) is the same as that in FIG.

[0173] After step S1131, the administrator client device 10 executes presentation of the summary (S3111). Specifically, the processor 12 displays screen P3110 (FIG. 17) on the display. The screen P3110 includes a display object A3110, an operation object B1110, and a field object F1111a. The operation object B 1110 and the field object F 1111a are the same as those in FIG.

[0174] The display object A 3110 is an object that displays chat history information included in the generated response data.

[0175] After step S3111, the administrator client device 10 executes a registration request (S3112). Specifically, when the administrator operates the operation object B 3110, the processor 12 transmits registration request data to the server 30. The registration request data includes, for example, the following information: Tag information entered in field object F1111a

[0176] After step S3112, the server 30 executes updating of the database (S3132). Specifically, the processor 32 adds a new record to the knowledge database (FIG. 15) in response to the registration request data. The following information is stored in each field of the new record: "Knowledge ID" field: New knowledge identification information "Summary" field: chat summary information obtained in step S3130 "Tag" field: tag information included in the registration request data

[0177] After step S3132, the server 30 executes a registration response (S1133) in the same manner as in FIG.

[0178] The following steps are the same as in FIG. Knowledge Usage Request (S1113) · Knowledge Use (S1134) ~ Knowledge Use Response (S1135) Completion notification (S1114)

[0179] (6-2-4) Summary of Modification 2 According to the second modification, a chat summary is generated by inputting the chat history into the trained model. The generated summary is registered in the knowledge database. This allows the administrator to increase the amount of information in the knowledge database without increasing the man-hours required to input information to the knowledge database. As a result, it is possible to further remove constraints on the improvement effect of the chat system using artificial intelligence.

[0180] According to this embodiment, the generated information (chat summary) may be presented to an administrator. In this case, the administrator, after viewing the presented summaries, issues an instruction to register only the desired summaries in the knowledge database. Only the generated information according to the administrator's instruction (i.e., the summaries that the administrator determines should be registered) is registered in the knowledge database. This makes it possible to prevent low-quality generated information (i.e., summaries that are not useful for processing chats) from being mixed into the knowledge database.

[0181] (6-3) Variation 3 A description will now be given of Modification 3. Modification 3 is an example in which knowledge is utilized by referring to an evaluation of the chat history.

[0182] (6-3-1) Overview of Modification 3 An overview of Modification 3 will be described below. Fig. 18 is an explanatory diagram of an overview of Modification 3.

[0183] As shown in FIG. 18, the administrator provides chat history information to the server 30 via the administrator client device 10.

[0184] The server 30 generates generated information (at least one of a question, an answer, and chat summary information) using the chat history information and the evaluation information. The server 30 registers the generated information in a knowledge database.

[0185] (6-3-2) Knowledge registration process A description will now be given of the process of knowledge registration in the modified example 3. Fig. 19 is a sequence diagram of the process of knowledge registration in the modified example 3.

[0186] The administrator client device 10 executes the creation request (S1110). Specifically, the processor 22 displays a screen P1110 (FIG. 8) on the display. When the administrator inputs any thread identification information into the field object F1110b, the processor 12 transmits creation request data to the server 30. The creation request data includes, for example, the following information: Thread identification information entered in field object F1110b

[0187] After step S1110, the server 30 executes generation (S4130). Specifically, the storage device 31 stores a generative model. A generative model is a trained model trained through machine learning. The generative model uses the evaluation information included in the chat history information as a parameter. The generative model is configured to output the generated information according to a combination of the chat history information and the evaluation information. More specifically, the generative model is adjusted so that the higher the evaluation information, the stronger the influence of the chat history information on the generated information.

[0188] In a first example of step S4130, the processor 32 refers to the chat history database (FIG. 6) to identify the chat history information and the evaluation information corresponding to the thread identification information included in the creation request data. Processor 32 provides the chat history information and the rating information to a generative model. The generation model outputs generated information according to the chat history information and the evaluation information, The generated information is at least one of a question, an answer, and chat summary information.

[0189] In the second example of step S4130, the storage device 31 stores two types of mask models (a first mask model and a second mask model).

[0190] The processor 32 selects a mask model (at least one of the first mask model and the second mask model) corresponding to the information on the mask type included in the generation request data. The processor 32 refers to the chat history database (FIG. 6) to identify the chat history information and the evaluation information corresponding to the thread identification information included in the creation request data. Processor 32 applies the chat history information to the selected mask model. The mask model outputs information in which mask target information has been removed from chat history information (hereinafter referred to as "mask input information"). The processor 32 provides the masked input information and the evaluation information to the generative model. The generative model outputs generated information according to a combination of the masked input information and the evaluation information, The generated information is at least one of a question, an answer, and chat summary information.

[0191] The following steps are the same as in FIG. Generated response (S1131) · Presentation of generated results (S1111) ~ Registration request (S1112) · Database update (S1132) ~ Registration response (S1133) Knowledge Usage Request (S1113) · Knowledge Use (S1134) ~ Knowledge Use Response (S1135) Completion notification (S1114)

[0192] (6-3-3) Summary of Variation 3 According to the third modification, the degree of influence that the chat history information has on the generated information is adjusted according to the evaluation information. This makes it easier for chat history information with high evaluation information to have an effect on the generated information, which further removes limitations on the improvement effect of the chat system using artificial intelligence.

[0193] (7) Other modifications Other modifications will now be described.

[0194] In this embodiment, the generation response (S1131) and the presentation of the generation result (S1111) through the registration request (S1112) can be omitted. In this case, the server 30 updates the knowledge database (FIG. 5) based on the result of the generation (S1130) without instructions from the administrator.

[0195] In the second modification, the generated response (S1131) and the presentation of the abstract (S3111) through the registration request (S3112) can be omitted. In this case, the server 30 updates the knowledge database (FIG. 15) based on the result of the generation (S3130) without instructions from the administrator.

[0196] In the knowledge registration process (FIG. 7), a pattern in which one pair of question and answer is generated has been shown, but the present embodiment is not limited to this. This embodiment is also applicable to the following examples. (First pattern) A pattern in which a combination of one question and multiple answers is generated (Second pattern) A pattern in which a combination of multiple questions and one answer is generated (Third pattern) A pattern in which a combination of multiple questions and multiple answers is generated In this case, the server 30 may generate a question and an answer in a pattern designated by the administrator or visitor from among the above four patterns (i.e., a pattern in which a pair of question and answer is generated, and the above first to third patterns).

[0197] In the chat process (FIG. 9), an example in which the server 30 generates answers to questions from visitors (a so-called chatbot example) has been shown, but the present embodiment is not limited to this. This embodiment is also applicable to an example in which a human (for example, an administrator) inputs the answer. In this case, the answer entered by the administrator is accepted in updating the database (S1230).

[0198] In the present embodiment, an example has been described in which knowledge category information is determined in response to an instruction from an administrator, but the scope of the present embodiment is not limited to this. This embodiment can also be applied to the following examples. Specifically, a knowledge category determination model is stored in the storage device 31. The knowledge category determination model is configured to output knowledge category information based on knowledge information. The knowledge category determination model is, for example, at least one of the following. A trained model trained using a combination of at least one of questions and answers, and knowledge category information as training data. A rule-based model that describes the correlation between at least one of questions and answers and knowledge category information The processor 32 inputs the knowledge information stored in the knowledge database (FIG. 5) to the knowledge category determination model. The knowledge category determination model outputs knowledge category information corresponding to the input knowledge information. The processor 32 stores the knowledge category information output from the knowledge category determination model in the "knowledge category" field of the knowledge database.

[0199] In the present embodiment, an example in which chat category information is determined in response to an instruction from an administrator has been described, but the scope of the present embodiment is not limited to this. This embodiment can also be applied to the following examples. Specifically, a chat category determination model is stored in the storage device 31. The chat category determination model is configured to output chat category information based on chat information. The chat category determination model is, for example, at least one of the following: A trained model trained using a combination of at least one of questions and answers, and chat category information as training data. A rule-based model describing correlations between at least one of questions and answers and chat category information The processor 32 inputs chat information stored in the chat database (FIG. 6) into the chat category determination model. The chat category determination model outputs chat category information corresponding to the input chat information. The processor 32 stores the chat category information output from the chat category determination model in the “chat category” field of the chat database.

[0200] The storage device 11 may be connected to the administrator client device 10 via a network NW. The storage device 21 may be connected to the visitor client device 20 via a network NW. The storage device 31 may be connected to the server 30 via a network NW.

[0201] Each step of the above information processing can be executed by either the administrator client device 10 or the server 30. For example, if the administrator client device 10 is capable of executing all the steps of the above-mentioned information processing, the administrator client device 10 functions as an information processing device operating in a stand-alone manner without transmitting requests to the server 30 .

[0202] Although the embodiment of the present invention has been described in detail above, the scope of the present invention is not limited to the above embodiment. Furthermore, the above embodiment can be improved or modified in various ways without departing from the spirit of the present invention. Furthermore, the above embodiment and the modified examples can be combined. [Explanation of symbols]

[0203] 1: Information processing system 10: Administrator client device 11:Storage device 12: Processor 13: Input / Output Interface 14: Communication interface 20: Visitor client device 21:Storage device 22: Processor 23: Input / Output Interface 24: Communication interface 30: Server 31:Storage device 32 : Processor 33: Input / Output Interface 34: Communication interface

Claims

1. A method for processing information of network services provided using a network, Computers The steps include obtaining at least one piece of input information: visitor input information entered by a visitor to the network service, and administrator input information entered by the administrator of the network service. The step of generating new generated information based on the input information is performed by inputting the aforementioned input information into a trained model. The generated information includes questions relating to the network service and answers to those questions. The step of presenting the generated information to the administrator is performed, In response to the administrator's instruction to register the presented generated information, the step of registering the generated information in the database is performed. Information processing methods.

2. The input information is at least one of a file, the result of scraping a website or intranet site, text, and chat history provided in the network service. The information processing method according to claim 1.

3. The generation step involves inputting one of the input pieces of information into the trained model to generate a plurality of the generated pieces of information. The information processing method according to claim 1 or claim 2.

4. A further step is performed to remove the information to be masked from the input information by providing the input information to a mask model corresponding to the rules defined by the administrator, prior to the generation step, The generation step involves inputting the input information from which the mask target information has been removed into the trained model. The information processing method according to claim 1 or claim 2.

5. Before the generation step, the step of removing the information to be masked from the input information is further performed by providing the input information to a mask model corresponding to different rules for each visitor defined by the visitor, The generation step involves inputting the input information from which the mask target information has been removed into the trained model. The information processing method according to claim 1 or claim 2.

6. After the registration step, the further step is to fine-tune the answer model that generates answers to questions using the generated information registered in the database. The information processing method according to claim 1 or claim 2.

7. After the registration step, the further step of generating an answer model that generates answers to questions using the generated information registered in the database as a training dataset is performed. The information processing method according to claim 1 or claim 2.

8. After the registration step, the step of generating an FAQ using the generated information registered in the database is further performed. The information processing method according to claim 1 or claim 2.

9. The input information is the chat history between the chatbot provided in the network service or between the administrator and the visitor, The generation step involves inputting the chat history and evaluation information for the chat history into the trained model to generate the generated information according to the combination of the chat history and the evaluation information. The information processing method according to claim 1.

10. The computer further performs the step of associating and storing the generated information with knowledge category information. The information processing method according to claim 1 or claim 2.

11. The computer further performs the step of associating and storing the generated information and tag information, The tag information is set by the administrator, or by using a tag model that generates at least one synonym for the question and the answer. The information processing method according to claim 1 or claim 2.

12. A computer, It functions as a means for acquiring at least one piece of input information: visitor input information entered by a visitor to a network service, and administrator input information entered by the administrator of the network service. By inputting the aforementioned input information into a trained model, it functions as a means for generating new generated information based on the aforementioned input information. The generated information includes questions relating to the network service and answers to those questions. The generated information is to be used as a means to present it to the administrator. In response to the administrator's instruction to register the presented generated information, the generated information is to be used as a means to register the generated information in a database. A program for that purpose.

13. A means for acquiring at least one piece of input information: visitor input information entered by a visitor to a network service, and administrator input information entered by the administrator of the network service, The system includes means for generating new generated information based on the input information by inputting the aforementioned input information into a trained model, The generated information includes questions relating to the network service and answers to those questions. The system includes means for presenting the generated information to the administrator, The system includes means for registering the generated information in a database in response to the administrator's instruction to register the generated information presented. Information processing device.

14. An information processing system comprising a client device and a server connected to the client device, The system includes means for acquiring at least one piece of input information: visitor input information entered by a visitor to a network service, and administrator input information entered by the administrator of the network service. The system includes means for generating new generated information based on the input information by inputting the aforementioned input information into a trained model, The generated information includes questions relating to the network service and answers to those questions. The system includes means for presenting the generated information to the administrator, The system includes means for registering the generated information in a database in response to the administrator's instruction to register the generated information presented. Information processing system.

Images