Search control system, search control method and program

The search control system addresses the cumbersome process of separate query input by generating search queries from chat interactions, allowing seamless retrieval of results through a language model integration.

JP7829023B1Active Publication Date: 2026-03-12RAKUTEN GROUP INC
View PDF 5 Cites 0 Cited by

Patent Information

Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Filing Date
2024-12-24
Publication Date
2026-03-12

AI Technical Summary

Technical Problem

Users must separately input search queries into a search system when using a language model for information retrieval, which is cumbersome.

Method used

A search control system that generates and outputs search queries seamlessly based on chat messages, using a language model to determine user intent and interact naturally, integrating with search systems to provide results.

Benefits of technology

Enables users to obtain search results naturally, as if chatting with a person, eliminating the need for separate query input.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007829023000001_ABST
    Figure 0007829023000001_ABST
Patent Text Reader

Abstract

To provide a search control system, a search control method, and a program that enable a user to seamlessly obtain search results from a search system with the feeling that he or she is chatting naturally with another person. [Solution] A prompt generation unit 46 generates a prompt for a language model based on a chat message received from a terminal 14. A prompt output unit 48 outputs the prompt to the language model. A display control unit 56 outputs a chat message based on a response from the language model to the prompt to the terminal 14. A first search intermediation unit 50a generates a search query based on a log of chat message exchanges with the terminal 14. The first search intermediation unit 50a outputs the search query to a first search system 12a. The display control unit 56 outputs search results received from the first search system 12a in response to the output of the search query to the first search system 12a to the terminal 14.
Need to check novelty before this filing date? Find Prior Art

Description

[Technical Field]

[0001] The present invention relates to a search control system, a search control method, and a program. [Background technology]

[0002] Patent Document 1 describes a technology that supports users in searching for products on an internet shopping site.

[0003] In recent years, technology utilizing large-scale language models (LLMs) has been attracting attention. As an example of such technology, Patent Document 2 describes a technology that uses a large-scale language model when generating an automated assistant response. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Laid-Open No. 2011-133926 [Patent Document 2] Special Publication No. 2024-521053 Summary of the Invention [Problem to be solved by the invention]

[0005] By repeatedly outputting prompts to a language model such as LLM and receiving responses from that language model, users can obtain information as if they were chatting naturally with a person.

[0006] However, in order to perform some kind of search based on information obtained through chat using a language model, the user must separately input a search query into the search system, which is cumbersome.

[0007] The present invention has been made in consideration of the above-mentioned problems, and one of its objectives is to provide a search control system, search control method, and program that allows users to obtain search results from a search system seamlessly, as if they were chatting naturally with a person. [Means for solving the problem]

[0008] The search control system of the present invention includes a prompt generation means for generating a prompt to a language model based on a chat message received from a terminal, a prompt output means for outputting the prompt to the language model, a chat message output means for outputting a chat message to the terminal based on a response from the language model to the prompt, a search query generation means for generating a search query based on a log of chat message exchanges with the terminal, a search query output means for outputting the search query to a search system, and a search result output means for outputting search results received from the search system in response to the output of the search query to the search system to the terminal.

[0009] The search control method of the present invention includes the steps of generating a prompt for a language model based on a chat message received from a terminal, outputting the prompt to the language model, outputting a chat message based on a response from the language model to the prompt to the terminal, generating a search query based on a log of chat message exchanges with the terminal, outputting the search query to a search system, and outputting search results received from the search system in response to the output of the search query to the search system to the terminal.

[0010] The program of the present invention is a program for causing a computer to execute the following steps: generating a prompt for a language model based on a chat message received from a terminal; outputting the prompt to the language model; outputting a chat message based on a response from the language model to the prompt to the terminal; generating a search query based on a log of chat message exchanges with the terminal; outputting the search query to a search system; and outputting search results received from the search system in response to the output of the search query to the search system to the terminal. [Brief explanation of the drawings]

[0011] [Figure 1] 1 is a diagram illustrating an example of a configuration of a chat system according to an embodiment of the present invention. [Figure 2] FIG. 10 is a diagram illustrating an example of a chat screen. [Figure 3] FIG. 10 is a diagram illustrating an example of a chat screen. [Figure 4] FIG. 2 is a functional block diagram showing an example of functions of the search control system according to one embodiment of the present invention. [Figure 5A] FIG. 2 is a flowchart showing an example of the flow of processing performed in the search control system according to one embodiment of the present invention. [Figure 5B] FIG. 2 is a flowchart showing an example of the flow of processing performed in the search control system according to one embodiment of the present invention. [Figure 6] FIG. 10 is a functional block diagram showing an example of functions of a search control system according to a modified example of an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0012] Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

[0013] Fig. 1 is a diagram showing an example of the configuration of a chat system 1 according to this embodiment. As shown in Fig. 1, the chat system 1 according to this embodiment includes a search control system 10, a search system 12 (a first search system 12a and a second search system 12b), and a terminal 14. The search control system 10, the first search system 12a, the second search system 12b, and the terminal 14 are connected to a computer network 16 such as the Internet.

[0014] The search control system 10 may be configured, for example, by a computer such as a server computer. The search control system 10 may include a processor 10a such as a CPU or GPU, a storage unit 10b such as a memory or a hard disk drive, and a communication unit 10c such as a network interface card. The search control system 10 may include multiple computers or may include a single computer.

[0015] The search system 12 (first search system 12a and second search system 12b) may be configured by a server computer including, for example, a CPU, a GPU, a memory, a communication interface, etc. Furthermore, the search system 12 may include multiple computers or may include a single computer.

[0016] The terminal 14 may be configured by, for example, a computer (for example, a personal computer, a smartphone, a tablet terminal, etc.) including a CPU, a memory, a communication interface, an input device, and a display.

[0017] In this embodiment, for example, chat messages are exchanged between the search control system 10 and the terminal 14. During the exchange, a search is executed in the search system 12 as appropriate. In the search system 12, for example, a search is performed for suggestions to the user (more specifically, for example, a search for generating options for products or services to be suggested for purchase to the user).

[0018] In this embodiment, for example, the search targets in a search by the first search system 12a and the search targets in a search by the second search system 12b are different. Furthermore, the search system 12 may execute a search for products or services as search targets. For example, the search targets in a search by the first search system 12a may be data indicating products available for sale in online shopping, while the search targets in a search by the second search system 12b may be data indicating travel packages.

[0019] 2 is a diagram showing an example of a chat screen 20 displayed on the display of the terminal 14 according to this embodiment. The chat screen 20 may be a screen that can be scrolled in a predetermined direction (for example, up and down).

[0020] On the chat screen 20 shown in FIG. 2, user message images 22 (22a, 22b, and 22c) showing chat messages input by users operating the terminals 14 are arranged.

[0021] 2, reply message images 24 (24a and 24b) are arranged to indicate chat messages sent from the search control system 10 to the terminal 14 in response to chat messages received from the terminal 14 by the search control system 10. The reply message images 24 may also include option images 26 indicating options selectable by the user. In the example of FIG. 2, the reply message image 24a includes option images 26a, 26b, 26c, and 26d. The reply message image 24b includes option images 26e, 26f, 26g, 26h, and 26i.

[0022] Furthermore, the chat screen 20 shown in FIG. 2 has an input form 28 arranged thereon for the user to input a chat message.

[0023] When a user enters a string of characters into the input form 28 and performs a predetermined operation (for example, pressing the Enter key or tapping the send button), a chat message including the string of characters entered into the input form 28 is sent from the terminal 14 to the search control system 10.

[0024] In addition, when the user performs an operation to select an optional image 26 (for example, a tap operation on the optional image 26), a chat message including the character string placed on the selected optional image 26 is sent from the terminal 14 to the search control system 10.

[0025] Furthermore, the chat screen 20 to which the user message image 22 containing the chat message transmitted from the terminal 14 to the search control system 10 has been added is displayed on the display of the terminal 14 .

[0026] The chat screen 20 shown in FIG. 2 shows a situation in which a chat message "Search for me" is sent from the terminal 14 to the search control system 10 as a result of a selection operation being performed on the option image 26i.

[0027] Then, in this embodiment, for example, as shown in Fig. 3, search results from a search executed in the search system 12 are transmitted from the search control system 10 to the terminal 14, and a chat screen 20 to which a search result image 30 showing the search result has been added is displayed on the display of the terminal 14. As shown in Fig. 3, the search results may indicate products or services. For example, as shown in Fig. 3, the search results may include an image of the product or service, the name of the product or service, and the price of the product or service.

[0028] 3, the display of the terminal 14 may display a chat screen 20 in which chat messages and search results are arranged in chronological order. For example, as shown in FIG. 3, the display may display a chat screen 20 in which chat messages and search results are arranged in chronological order from top to bottom.

[0029] In this embodiment, for example, when the user performs a predetermined operation (for example, a tap operation) on the search result shown in the search result image 30, the screen displayed on the display of the terminal 14 transitions to a purchase page for purchasing the product or service shown in the search result. Then, the user can purchase the product or service via the purchase page.

[0030] In this way, in this embodiment, the user can seamlessly obtain search results from the search system 12, as if he or she were chatting naturally with another person.

[0031] The functions of the search control system 10 according to this embodiment and the processes executed by the search control system 10 will be further described below.

[0032] Fig. 4 is a functional block diagram showing an example of functions implemented in the search control system 10 according to this embodiment. Note that the search control system 10 according to this embodiment does not need to implement all of the functions shown in Fig. 4, and functions other than those shown in Fig. 4 may also be implemented.

[0033] As shown in FIG. 4, the search control system 10 according to this embodiment functionally includes, for example, a log storage unit 40, a chat message receiving unit 42, an intent determination unit 44, a prompt generation unit 46, a prompt output unit 48, a search intermediation unit 50 (a first search intermediation unit 50a and a second search intermediation unit 50b), a general dialogue unit 52, a chat message generation unit 54, and a display control unit 56. The log storage unit 40 is implemented primarily in the storage unit 10b. The chat message receiving unit 42 is implemented primarily in the communication unit 10c. The intent determination unit 44, the prompt generation unit 46, the prompt output unit 48, the general dialogue unit 52, and the chat message generation unit 54 are implemented primarily in the processor 10a. The first search intermediation unit 50a, the second search intermediation unit 50b, and the display control unit 56 are implemented primarily in the processor 10a and the communication unit 10c.

[0034] The above functions may be implemented by executing a program containing instructions corresponding to the above functions, which is installed in the search control system 10, which is a computer. The program may also be supplied to the search control system 10 via a computer-readable information storage medium such as an optical disk, a magnetic disk, a magnetic tape, or a magneto-optical disk, or via the Internet, for example.

[0035] In this embodiment, the log storage unit 40 stores, for example, a log of chat message exchanges between the search control system 10 and the terminal 14. For example, the log storage unit 40 may store logs of chat messages that the search control system 10 has received from the terminal 14 and chat messages that the search control system 10 has output to the terminal 14. Here, for example, the log storage unit 40 may accumulate chat messages associated with a timestamp indicating the date and time when the chat message was received or output, and a flag indicating whether the chat message was received by the search control system 10 or output by the search control system 10.

[0036] In this embodiment, the chat message receiving unit 42 receives, for example, a chat message from the terminal 14. In addition, in this embodiment, the chat message receiving unit 42 stores the chat message in the log storage unit 40, for example.

[0037] In this embodiment, the intent determination unit 44 determines the user's intent based on chat messages received from the terminal 14, using existing intent determination techniques such as natural language processing (NLP) models, large-scale language models (LLMs), and knowledge graphs.

[0038] For example, the intent determination unit 44 may include a trained language model (e.g., LLM) for intent determination. Hereinafter, this language model will be referred to as the intent determination LLM. Then, the intent determination unit 44 may output to the intent determination LLM a prompt including, for example, a template such as "Please refer to the log located at the following location to determine your intent," and a character string indicating the location of the log storage unit 40 (e.g., a URL). Then, the intent determination unit 44 may determine, as the user's intent, a response output from the intent determination LLM in response to the output of the prompt. In this way, the intent determination unit 44 may determine the user's intent based on the log stored in the log storage unit 40.

[0039] Here, for example, the intent determination unit 44 may determine the user's intent based on at least one of the logs of chat messages received from the terminal 14 and the logs of chat messages output to the terminal 14, which are stored in the log storage unit 40.

[0040] Furthermore, the user's intent may be determined based on a group of chat messages in which each chat message is weighted so that the more recently a chat message was received or output, the heavier the weight.

[0041] Furthermore, the intent determination unit 44 may determine the user's intent based on the latest chat message received from the terminal 14. For example, the intent determination unit 44 may output to the intent determination LLM a template such as "Please determine your intent based on the following message," and a prompt including the latest chat message received from the terminal 14. Then, the intent determination unit 44 may determine, as the user's intent, a response output from the intent determination LLM in response to the output of the prompt.

[0042] In this embodiment, the prompt generation unit 46 generates a prompt for a language model (e.g., LLM) based on, for example, a chat message received from the terminal 14. Here, for example, the prompt generation unit 46 may generate a prompt including a chat message received by the chat message receiving unit 42.

[0043] In this embodiment, the prompt output unit 48 outputs the prompt generated by the prompt generation unit 46 to the language model, for example.

[0044] Here, a trained language model (e.g., an LLM) may be included in each of the first search mediation unit 50a, the second search mediation unit 50b, and the general dialogue unit 52. Hereinafter, the language model included in the first search mediation unit 50a will be referred to as the first LLM, the language model included in the second search mediation unit 50b will be referred to as the second LLM, and the language model included in the general dialogue unit 52 will be referred to as the third LLM.

[0045] Then, the prompt output unit 48 may output the prompt generated by the prompt generation unit 46 to any one of the first LLM, the second LLM, or the third LLM, for example.

[0046] Here, the intent determination unit 44 may determine a language model (hereinafter referred to as a target model) to which a prompt is to be output from among a plurality of language models, based on the log stored in the log storage unit 40. For example, the intent determination unit 44 may determine a user's intent based on the log stored in the log storage unit 40, and determine a target model from among a plurality of language models based on the determined intent. Here, for example, the intent determination unit 44 may store data associating intents with language models. Then, the intent determination unit 44 may determine, as the target model, a language model associated with the intent determined by the intent determination unit 44 in the data. Then, the prompt output unit 48 may output a prompt to the determined target model.

[0047] Specifically, for example, if it is determined that the user's intent is to purchase a product (e.g., purchase a gift for a partner), the intent determination unit 44 may determine the first LLM as the target model. Alternatively, for example, if it is determined that the user's intent is to travel (e.g., travel with a partner), the intent determination unit 44 may determine the second LLM as the target model. Alternatively, if the user's intent is not one of the above, the intent determination unit 44 may determine the third LLM as the target model. A prompt may then be output to the target model determined in this manner.

[0048] Furthermore, in this embodiment, the intent determination unit 44 may repeatedly execute the process of determining a target model from among a plurality of language models based on the log stored in the log storage unit 40. For example, the intent determination unit 44 may execute the process of determining a target model every time a chat message is received from the terminal 14. Then, the prompt output unit 48 may output a prompt based on the chat message received from the terminal 14 to the target model when the chat message was received.

[0049] In this embodiment, the search mediation units 50 (for example, the first search mediation unit 50a and the second search mediation unit 50b) generate responses to the received prompts. In this embodiment, knowledge about the search target (for example, products for sale in online shopping) in a search using the first search system 12a is pre-registered (learned) in the first LLM, which is a language model included in the first search mediation unit 50a. In this embodiment, knowledge about the search target (for example, travel packages) in a search using the second search system 12b is pre-registered (learned) in the second LLM, which is a language model included in the second search mediation unit 50b. In this embodiment, both the first LLM and the second LLM are capable of referencing logs stored in the log storage unit 40.

[0050] The search intermediation unit 50 may determine, based on the log stored in the log storage unit 40, whether or not sufficient information (context) for executing a search by the search system 12 has been obtained.

[0051] For example, if the target model is the first LLM, the prompt generation unit 46 may generate a prompt including a template such as "Please refer to the log at the following location to determine whether a product search is possible," and a character string indicating the location (e.g., a URL) of the log storage unit 40. If the target model is the first LLM, the prompt output unit 48 may output the prompt to the first LLM. If the answer output from the first LLM to the prompt is affirmative (e.g., "Yes, it is possible"), the first search intermediation unit 50a may determine that sufficient information (context) has been obtained to perform a search using the first search system 12a.

[0052] Furthermore, when the target model is the second LLM, the prompt generation unit 46 may generate a prompt including a template such as "Please refer to the log at the following location to determine whether a travel package search is possible," and a character string indicating the location (e.g., a URL) of the log storage unit 40. The prompt output unit 48 may output the prompt to the second LLM. Then, when the answer output from the second LLM to the prompt is affirmative (e.g., "Yes, it is possible"), the second search intermediation unit 50b may determine that sufficient information (context) has been obtained to perform a search using the second search system 12b.

[0053] Furthermore, for example, the search intermediation unit 50 may determine whether at least one predetermined attribute (e.g., gender, age group, address, hobbies, number of people in household, etc.) can be extracted from the logs stored in the log storage unit 40. If it is determined that the attribute can be extracted, the search intermediation unit 50 may determine that sufficient information (context) has been obtained to execute a search by the search system 12. Note that the attribute determined by the first search intermediation unit 50a and the attribute determined by the second search intermediation unit 50b may be the same or different.

[0054] Then, as will be described later, if it is determined that sufficient information (context) has been obtained to perform a search using the search system 12, a predetermined chat message indicating that a search can be performed (for example, a chat message including an optional image 26i with the string "Search for me" placed on it) may be sent to the terminal 14.

[0055] In this embodiment, the search intermediation unit 50 may generate a search query based on a log of chat message exchanges between the search control system 10 and the terminal 14. The search query may be a keyword used to search for a product or service. For example, the search intermediation unit 50 may generate a search query in response to a given search execution condition being satisfied. The search execution condition may be, for example, a condition that a chat message received by the chat message receiving unit 42 contains a predetermined character string (e.g., "Search for me"). The predetermined character string does not have to be the character string "Search for me" but may be, for example, the Japanese character string "Search for me." The search execution condition may also be a condition that it is determined that sufficient information (context) has been obtained to execute a search by the search system 12.

[0056] Here, the prompt generation unit 46 may generate a search query generation prompt that instructs generation of a search query based on the log stored in the log storage unit 40, for example, in response to a chat message received by the chat message receiving unit 42 satisfying a given search execution condition, and the prompt output unit 48 may output the search query generation prompt to the language model. Then, in response to outputting the search query generation prompt to the language model, the search intermediation unit 50 may generate a search query based on a response output from the language model.

[0057] For example, if the target model is the first LLM, a search query generation prompt generated by the prompt generation unit 46 may be output to the first LLM. Then, in response to the output of the search query generation prompt to the first LLM, the first search intermediation unit 50a may generate a search query based on a response output from the first LLM.

[0058] For example, the prompt generation unit 46 may generate a prompt including a template such as "Please refer to the log located at the following location and generate a group of Japanese keywords suitable for a product search in one sentence," and a character string indicating the location (e.g., a URL) of the log storage unit 40, and the prompt output unit 48 may output the prompt to the first LLM. Alternatively, the prompt generation unit 46 may generate a prompt including the contents of the log to be referenced and an instruction to generate a search query, and the prompt output unit 48 may output the prompt to the first LLM. The first search intermediation unit 50a may then acquire, as a search query, a response output from the first LLM in response to the prompt.

[0059] Also, for example, if the target model is a second LLM, a search query generation prompt generated by the prompt generation unit 46 may be output to the second LLM. Then, in response to the output of the search query generation prompt to the second LLM, the second search intermediation unit 50b may generate a search query based on a response output from the second LLM.

[0060] For example, the prompt generation unit 46 may generate a prompt including a template such as "Please refer to the log at the following location and generate a single sentence of Japanese keywords suitable for searching for travel packages," and a character string indicating the location (e.g., a URL) of the log storage unit 40, and the prompt output unit 48 may output the prompt to the second LLM. Alternatively, the prompt generation unit 46 may generate a prompt including the contents of the log to be referenced and an instruction to generate a search query, and the prompt output unit 48 may output the prompt to the second LLM. The second search intermediation unit 50b may then obtain, as a search query, a response output from the second LLM in response to the prompt.

[0061] The search intermediary unit 50 may generate a search query based on a chat message received from the terminal 14 (for example, the most recent chat message received from the terminal 14).

[0062] Hereinafter, a search query generated by the search intermediation unit 50 will be referred to as a generated search query.

[0063] Then, in this embodiment, for example, the search intermediation unit 50 outputs the generated search query to a search system 12 (e.g., a search system 12 associated with the search intermediation unit 50). Then, in this embodiment, for example, the search intermediation unit 50 accepts search results output from the search system 12 in response to the output of the generated search query to the search system 12.

[0064] For example, the first search intermediation unit 50a outputs the generated search query generated by the first search intermediation unit 50a to the first search system 12a. Then, in this embodiment, for example, the first search intermediation unit 50a accepts search results output from the first search system 12a in response to the output of the generated search query to the first search system 12a.

[0065] In addition, in this embodiment, for example, the second search intermediation unit 50b outputs the generated search query generated by the second search intermediation unit 50b to the second search system 12b. Then, in this embodiment, for example, the second search intermediation unit 50b accepts search results output from the second search system 12b in response to the output of the generated search query to the second search system 12b.

[0066] In this embodiment, when a search is performed in the search system 12, the prompt generation unit 46 may generate a prompt including the search results of the search in the search system 12, and the prompt output unit 48 may output the prompt to the language model. The prompt may include a chat message accepted by the chat message acceptance unit 42. The prompt may also include an instruction to refer to a log stored in the log storage unit 40. The search mediation unit 50 may then obtain an answer output from the language model in response to the prompt.

[0067] For example, if the target model is the first LLM and a search is performed in the first search system 12a, the prompt generator 46 may generate a prompt including the search results of the search in the first search system 12a, and the prompt output unit 48 may output the prompt to the first LLM. Then, the first search mediation unit 50a may obtain an answer to the prompt output from the first LLM.

[0068] For example, if the target model is the second LLM and a search is performed in the second search system 12b, the prompt generator 46 may generate a prompt including the search results of the search in the second search system 12b, and the prompt output unit 48 may output the prompt to the second LLM. Then, the second search mediation unit 50b may obtain an answer to the prompt output from the second LLM.

[0069] On the other hand, if a search is not performed in the search system 12, the prompt generation unit 46 may generate a prompt including the chat message received by the chat message reception unit 42, and the prompt output unit 48 may output the prompt to the language model. The prompt may include an instruction to refer to a log stored in the log storage unit 40. The search mediation unit 50 may then obtain an answer output from the language model in response to the prompt.

[0070] For example, if the target model is the first LLM and a search has not been performed in the first search system 12a, the prompt generation unit 46 may generate a prompt including the chat message received by the chat message reception unit 42, and the prompt output unit 48 may output the prompt to the first LLM. Then, the first search mediation unit 50a may obtain a response to the prompt output from the first LLM.

[0071] Furthermore, for example, if the target model is the second LLM and a search has not been performed in the second search system 12b, the prompt generation unit 46 may generate a prompt including the chat message received by the chat message reception unit 42, and the prompt output unit 48 may output the prompt to the second LLM. Then, the second search intermediation unit 50b may obtain a response to the prompt output from the second LLM.

[0072] In this embodiment, for example, the general dialogue unit 52 generates a response to the received prompt. In this embodiment, it is assumed that the third LLM can refer to the log stored in the log storage unit 40.

[0073] For example, when the target model is the third LLM, the prompt generation unit 46 may generate a prompt including the chat message received by the chat message reception unit 42, and the prompt output unit 48 may output the prompt to the third LLM. Note that the prompt may include an instruction to refer to a log stored in the log storage unit 40. Then, the general dialogue unit 52 may acquire a response to the prompt output from the third LLM.

[0074] In this embodiment, the chat message generating unit 54 generates chat messages to be output to the terminal 14, for example.

[0075] For example, when the target model is the first LLM, the chat message generation unit 54 may generate a chat message including an answer generated by the first LLM. Here, when the first search intermediation unit 50a receives search results output from the first search system 12a, the chat message generation unit 54 may generate a chat message including the search results received by the first search intermediation unit 50a from the first search system 12a and the answer generated by the first LLM.

[0076] Furthermore, for example, when the target model is the second LLM, the chat message generation unit 54 may generate a chat message including a response generated by the second LLM. Here, when the second search intermediation unit 50b receives search results output from the second search system 12b, the chat message generation unit 54 may generate a chat message including the search results received by the second search intermediation unit 50b from the second search system 12b and the response generated by the second LLM.

[0077] Furthermore, for example, when the target model is the third LLM, the chat message generating unit 54 may generate a chat message including a response generated by the third LLM.

[0078] In this embodiment, the chat message generation unit 54 may generate a predetermined chat message indicating that a search is executable when a given search executable condition is satisfied. For example, the chat message generation unit 54 may generate a chat message including an option image 26i in which the character string "Search for me" is placed when a given search executable condition is satisfied. Here, the search executable condition may be, for example, a condition that it is determined that sufficient information (context) for executing a search has been obtained.

[0079] Furthermore, a prompt including an instruction to generate options to be presented to the user may be output to the first, second, or third LLM, and the chat message generator 54 may generate a chat message including option images 26 associated with one or more options output from these LLMs in response to the output of the prompt.

[0080] In this embodiment, the display control unit 56 displays, for example, a chat screen 20 on the display of the terminal 14.

[0081] For example, in response to the chat message receiving unit 42 receiving a chat message from the terminal 14, the display control unit 56 may generate a chat screen 20 to which the chat message has been added and transmit the chat screen 20 to the terminal 14. Then, in response to receiving the chat screen 20, the terminal 14 may display the chat screen 20 on the display of the terminal 14.

[0082] Furthermore, for example, in response to the chat message generation unit 54 generating a chat message, the display control unit 56 may generate a chat screen 20 to which the chat message has been added, and transmit the chat screen 20 to the terminal 14. As described above, the chat message may include a search result (e.g., a search result image 30) received from the search system 12. Note that the chat message may include a portion of the search result received from the search system 12 (e.g., a search result image 30 showing a portion of the search result received from the search system 12). Then, in response to receiving the chat screen 20, the terminal 14 may display the chat screen 20 on the display of the terminal 14.

[0083] In this way, the display control unit 56 according to this embodiment serves as a chat message output unit that outputs a chat message based on a response from a language model to a prompt to the terminal 14. The display control unit 56 according to this embodiment also serves as a search result output unit that outputs search results received from the search system 12 to the terminal 14 in response to output of a generated search query to the search system 12.

[0084] In this embodiment, the search intermediary unit 50 may output the generated search query to the search system 12 in response to receiving a search request in response to the output of a predetermined chat message indicating that a search can be performed (for example, receiving a chat message "Search for me" in response to the output of a chat message including an optional image 26i in which the string "Search for me" is placed).

[0085] For example, in response to the first search intermediation unit 50a determining that sufficient information (context) for performing a search has been obtained, the chat message generation unit 54 may generate a chat message including an option image 26i in which the character string "Search for me" is placed. Then, the display control unit 56 may transmit a chat screen 20 including the option image 26i to the terminal 14. Then, in response to receiving a chat message "Search for me" in response to the transmission of the chat screen 20, the first search intermediation unit 50a may generate a generated search query and output the generated generated search query to the first search system 12a.

[0086] Furthermore, for example, in response to the second search intermediation unit 50b determining that sufficient information (context) for executing a search has been obtained, the chat message generation unit 54 may generate a chat message including an option image 26i in which the character string "Search for me" is placed. Then, the display control unit 56 may transmit a chat screen 20 including the option image 26i to the terminal 14. Then, in response to receiving a chat message "Search for me" in response to the transmission of the chat screen 20, the second search intermediation unit 50b may generate a generated search query and output the generated generated search query to the second search system 12b.

[0087] It is not necessary for the search intermediation unit 50 to generate a generated search query and output the generated search query to the search system 12 in response to receiving a search request in response to the output of a predetermined chat message indicating that a search is possible. Even in a situation where a predetermined chat message indicating that a search is possible has not been output, the search intermediation unit 50 may generate a generated search query in response to receiving a search request and output the generated search query to the search system 12. For example, in response to the chat message receiving unit 42 receiving a message including a predetermined character string (for example, the character string "search for" in Japanese), the search intermediation unit 50 may generate a generated search query and output the generated search query to the search system 12.

[0088] As described above, in this embodiment, the multiple language models (in the above example, the first LLM, the second LLM, and the third LLM) are each either a language model associated with the search system 12 or a language model not associated with the search system 12. In the above example, the first LLM and the second LLM are language models associated with the search system 12, and the third LLM is a language model not associated with the search system 12.

[0089] As described above, in this embodiment, each of the multiple language models is associated with a search system 12 having a different search target. In the above example, the first LLM is associated with the first search system 12a, and the second LLM is associated with the second search system 12b.

[0090] In this embodiment, for example, when the target model is a language model associated with the search system 12 and the above-mentioned search execution conditions are satisfied, a generated search query is output to the search system 12 associated with the target model.

[0091] For example, when the target model is the first LLM and the above-described search execution conditions are satisfied, the first search intermediation unit 50a outputs a generated search query to the first search system 12a. Then, the display control unit 56 outputs to the terminal 14 the search results for the search target of the first search system 12a, which are received from the first search system 12a in response to the output of the generated search query to the first search system 12a.

[0092] Furthermore, in response to the above-described search execution conditions being satisfied when the target model is the second LLM, the second search intermediation unit 50b outputs a generated search query to the second search system 12b. Then, the display control unit 56 outputs to the terminal 14 the search results for the search target of the second search system 12b, which are received from the second search system 12b in response to the output of the generated search query to the second search system 12b.

[0093] Furthermore, as described above, all of the multiple language models (in the above example, the first LLM, the second LLM, and the third LLM) may be able to refer to the logs stored in the log storage unit 40. The logs stored in the log storage unit 40 may be logs of chat message exchanges via the multiple language models.

[0094] In this embodiment, the prompt generation unit 46 may store the prompt generated by the prompt generation unit 46 in the log storage unit 40. In addition, the display control unit 56 may store the chat screen 20 generated by the display control unit 56 in the log storage unit 40.

[0095] In addition, in this embodiment, the first search intermediation unit 50a may store the generated search query generated by the first search intermediation unit 50a in the log storage unit 40. In addition, the second search intermediation unit 50b may store the generated search query generated by the second search intermediation unit 50b in the log storage unit 40.

[0096] Furthermore, in this embodiment, the log stored in the log storage unit 40 may include search results from searches performed by the search system 12. For example, the first search intermediation unit 50a may store the search results that the first search intermediation unit 50a receives from the first search system 12a in the log storage unit 40. For example, the second search intermediation unit 50b may store the search results that the second search intermediation unit 50b receives from the second search system 12b in the log storage unit 40.

[0097] The various data stored in the log storage unit 40 in this manner may be referenced in the various processes described above.

[0098] Here, an example of the flow of processing performed by the search control system 10 according to this embodiment will be described with reference to the flow diagrams shown in FIGS. 5A and 5B.

[0099] When the chat message receiving unit 42 receives a chat message from the terminal 14 (S101), the display control unit 56 generates a chat screen 20 including the chat message and transmits the chat screen 20 to the terminal 14 (S102). In response to receiving the chat screen 20, the terminal 14 updates the chat screen 20 displayed on the display to the received chat screen 20.

[0100] Then, the intent determination unit 44 determines the user's intent (S103), and then determines a target model based on the intent determined in the process shown in S103 (S104).

[0101] Then, the prompt generating unit 46 determines whether or not a search query needs to be generated based on the chat message received in the process shown in S101 (S105).

[0102] If it is determined that generation of a generated search query is unnecessary (S105: N), the prompt generation unit 46 generates a prompt including the chat message received in the process shown in S101 (S106), and the prompt output unit 48 outputs the prompt generated in the process shown in S106 to the target model (S107).

[0103] Then, the chat message generating unit 54 generates a chat message based on the response that is output from the target model in response to the prompt that was output to the target model in the process shown in S107 (S108).

[0104] Then, the display control unit 56 generates a chat screen 20 including the chat message generated in the process shown in S108 and transmits the chat screen 20 to the terminal 14 (S109), and the process shown in this process example ends. In response to receiving the chat screen 20, the terminal 14 updates the chat screen 20 displayed on the display to the received chat screen 20.

[0105] If it is determined in the process shown in S105 that generation of a search query is necessary (S105: Y), the prompt generation unit 46 generates a search query generation prompt (S110), and the prompt output unit 48 outputs the search query generation prompt generated in the process shown in S110 to the target model (S111).

[0106] Then, the search intermediary unit 50 including the target model acquires the generated search query that is output from the target model in response to the search query generation prompt output to the target model in the process shown in S111 (S112).

[0107] Then, the search intermediation unit 50 outputs the generated search query obtained by the process shown in S112 to the search system 12 associated with the search intermediation unit 50 (S113), and the search intermediation unit 50 accepts search results output from the search system 12 in response to the output of the generated search query to the search system 12 (S114).

[0108] Then, the prompt generating unit 46 generates a prompt including the search result received in the process shown in S114 (S115), and the prompt output unit 48 outputs the prompt generated in the process shown in S115 to the target model (S116).

[0109] Then, the chat message generation unit 54 generates a chat message based on the search result received in the process shown in S114 and the response that is output from the target model in response to the prompt output to the target model in the process shown in S116 (S117).

[0110] Then, the display control unit 56 generates a chat screen 20 including the chat message generated in the process shown in S117 and transmits the chat screen 20 to the terminal 14 (S118), and the process shown in this process example ends. In response to receiving the chat screen 20, the terminal 14 updates the chat screen 20 displayed on the display to the received chat screen 20.

[0111] The processes shown in S101 to S118 are executed every time the chat message receiving unit 42 receives a chat message.

[0112] According to this embodiment, as described above, the user can obtain search results from the search system 12 seamlessly via the chat screen 20, with the feeling that he or she is chatting naturally with another person.

[0113] In this embodiment, all of the multiple language models (the 1st LLM, 2nd LLM, and 3rd LLM in the above example) may be pre-configured to generate answers in a common communication style.

[0114] For example, a prompt such as "Communication style: · Use emojis when appropriate and maintain a light and positive tone. · Express excitement about a positive user experience. · Use the first person to create a more personal impression (e.g., "I recommend..."). · Offer non-authoritative advice and suggestions and act as a friendly assistant" may be pre-printed for all LLMs 1, 2, and 3.

[0115] In this way, it is possible to prevent the user from feeling uncomfortable even when chatting using a plurality of language models.

[0116] Furthermore, in this embodiment, as described above, the number of language models used to generate chat messages does not need to be three (in the above example, the first LLM, the second LLM, and the third LLM), nor does it need to be multiple.

[0117] For example, as shown in Figure 6, the search control system 10 may not include the first search mediation unit 50a, the second search mediation unit 50b, and the general dialogue unit 52, but may instead include a comprehensive search mediation unit 60. The comprehensive search mediation unit 60 may include a trained language model (e.g., an LLM). Hereinafter, the language model included in the comprehensive search mediation unit 60 will be referred to as the comprehensive LLM.

[0118] A plurality of templates may be stored in advance in the prompt generation unit 46. The intent determination unit 44 may then repeatedly execute a process of determining a target template from among the plurality of templates based on the logs stored in the log storage unit 40. For example, the intent determination unit 44 may execute a process of determining a target template each time a chat message is received from the terminal 14.

[0119] For example, the intent determination unit 44 may store data associating intents with templates, and may determine, as the target template, a template associated with the intent determined by the intent determination unit 44 in the stored data.

[0120] The prompt generation unit 46 may then generate a prompt based on the chat message received from the terminal 14 and the target template when the chat message is received. For example, the prompt generation unit 46 may generate a prompt that includes the chat message received by the chat message receiving unit 42 and the target template. The prompt output unit 48 may then output the prompt to the integrated LLM.

[0121] Furthermore, each of the multiple templates may be either a template associated with the search system 12 or a template not associated with the search system 12. Furthermore, each of the multiple templates may be associated with a search system 12 having a different search target.

[0122] For example, the plurality of templates may be a first template associated with the first search system 12a, a second template associated with the second search system 12b, and a third template not associated with any search system 12. Here, the first template may include information about the first search system 12a (e.g., the URL of the first search system 12a), and the second template may include information about the second search system 12b (e.g., the URL of the second search system 12b).

[0123] Then, in response to the target template being a template associated with the search system 12 and satisfying the above-mentioned search execution conditions, a generated search query may be output to the search system 12 associated with the target template. Then, in response to the output of the generated search query to the search system 12, search results for the search target of the search system 12 received from the search system 12 may be output to the terminal 14.

[0124] For example, when the target template is the first template and the above-described search execution conditions are satisfied, the comprehensive search intermediation unit 60 may output a generated search query to the first search system 12a. Then, the display control unit 56 may output to the terminal 14 search results for the search target of the first search system 12a, which are received from the first search system 12a in response to the output of the generated search query to the first search system 12a.

[0125] Furthermore, in a situation where the target template is the second LLM and the above-described search execution conditions are satisfied, the comprehensive search intermediation unit 60 may output a generated search query to the second search system 12b. Then, in response to the output of the generated search query to the second search system 12b, the display control unit 56 may output to the terminal 14 the search results for the search target of the second search system 12b received from the second search system 12b.

[0126] The present invention is not limited to the above-described embodiment.

[0127] For example, the language model according to this embodiment does not have to be a large-scale language model (LLM). For example, the language model according to this embodiment may be an n-gram model, a recurrent neural network (RNN), a Seq2Seq model, a word embedding model, or the like.

[0128] Furthermore, the search targets in the search system 12 according to this embodiment are not limited to products and services.

[0129] Furthermore, the number of search systems 12 included in the chat system 1 according to this embodiment does not need to be two.

[0130] Furthermore, the search intermediation units 50 and search systems 12 do not have to be associated one-to-one. For example, the first search intermediation unit 50a may be associated with multiple search systems 12. Furthermore, the second search intermediation unit 50b may be associated with multiple search systems 12.

[0131] Furthermore, at least some of the trained language models described above may not be included in the search control system 10, but may be included in an external system that can communicate with the search control system 10.

[0132] Furthermore, the specific character strings and numerical values ​​described above and the specific character strings and numerical values ​​in the drawings are examples, and the present invention is not limited to these character strings and numerical values.

[0133] Furthermore, for example, the search control system can also be configured as follows. [1] prompt generation means for generating prompts for a language model based on chat messages received from the terminal; prompt output means for outputting the prompt to the language model; a chat message output means for outputting to the terminal a chat message based on a response from the language model to the prompt; a search query generation means for generating a search query based on a log of chat message exchanges with the terminal; a search query output means for outputting the search query to a search system; a search result output means for outputting search results received from the search system to the terminal in response to the search query being output to the search system; A search control system including: [2] the search query generation means generates a search query generation prompt that instructs generation of the search query based on the log; the search query generation means generates the search query based on a response output from a language model in response to output of the search query generation prompt to the language model; [1] The search control system according to [1]. [3] a determination unit that repeatedly executes a process of determining a target model from among a plurality of language models based on the log; the prompt output means outputs a prompt based on a chat message received from the terminal to the target model when the chat message is received; the log is a log of chat message exchanges via the plurality of language models; [1] or [2]. The search control system according to [1] or [2]. [4] the determination means executes a process of determining the target model every time a chat message is received from the terminal; [3] A search control system according to the present invention. [5] each of the plurality of language models is either a language model associated with a search system or a language model not associated with a search system; the search query output means outputs the search query to a search system associated with the target model in response to a given condition being satisfied in a situation where the target model is a language model associated with a search engine; [3] or [4], the search control system. [6] the plurality of language models are associated with search systems having different search targets; the search query output means outputs the search query to a search system associated with the target model in response to a given condition being satisfied; the search result output means outputs to the terminal a search result for a search target of the search system, the search result being received from the search system in response to the output of the search query to the search system; [3] to [5]. A search control system according to any one of [3] to [5]. [7] all of the plurality of language models are able to refer to the log; [3] to [6]. A search control system according to any one of [3] to [6]. [8] all of the plurality of language models are pre-configured to generate answers in a common communication style; [3] to [7]. A search control system according to any one of [3] to [7]. [9] a determination unit that repeatedly executes a process of determining a target template from among a plurality of templates based on the log; the prompt output means outputs to the language model a prompt based on a chat message received from the terminal and the target template when the chat message is received. [1] or [2]. The search control system according to [1] or [2].

[10] the determination means executes a process of determining the target template every time a chat message is accepted from the terminal. [9] The search control system described in [9].

[11] each of the plurality of templates is either a template associated with a search system or a template not associated with a search system; the search query output means outputs the search query to the search system associated with the target template in response to a given condition being satisfied in a situation where the target template is a template associated with the search system; [9] or

[10] .

[12] each of the plurality of templates is associated with a search system having a search target different from the others; the search query output means outputs the search query to a search system associated with the target template in response to a given condition being satisfied; the search result output means outputs to the terminal a search result for a search target of the search system, the search result being received from the search system in response to the output of the search query to the search system; [9] to

[11] . A search control system according to any one of [9] to

[11] .

[13] the chat message output means outputs a predetermined chat message indicating that a search is possible when a given condition is satisfied; the search query output means outputs the search query to the search system in response to a search request received in response to the output of the predetermined chat message; [1] A search control system according to any one of [1] to

[12] .

[14] the search query output means outputs the search query to the search system in response to reception of a chat message indicating a request for a search for the output of the predetermined chat message;

[13] The search control system described in

[13] .

[15] The log includes search results from the search. [1] to

[14] .

[16] The search system executes a search targeting a product or service. [1] to

[15] .

[17] The search query is a keyword used to search for a product or service. [1] to

[16] . A search control system.

[18] the search results indicate products or services; [1] to

[17] .

[19] The terminal further includes: The terminal includes a display means for displaying a screen on which chat messages and the search results are arranged in chronological order. [1] to

[18] . A search control system. [Explanation of symbols]

[0134] 1 Chat system, 10 Search control system, 10a Processor, 10b Memory unit, 10c Communication unit, 12 Search system, 12a First search system, 12b Second search system, 14 Terminal, 16 Computer network, 20 Chat screen, 22, 22a, 22b, 22c User message image, 24, 24a, 24b Reply message image, 26, 26a, 26b, 26c, 26d, 26e, 26f, 26g, 26h, 26i Option image, 28 Input form, 30 Search result image, 40 Log memory unit, 42 Chat message reception unit, 44 Intent determination unit, 46 Prompt generation unit, 48 Prompt output unit, 50 Search mediation unit, 50a First search mediation unit, 50b Second search mediation unit, 52 General dialogue unit, 54 Chat message generation unit, 56 Display control unit, 60 General search intermediary department.

Claims

1. A determination means for repeatedly executing a process of determining a target model from a plurality of language models, each of which is either a language model associated with a search system or a language model not associated with a search system, based on a log of chat message exchanges between the terminal and the target model; a prompt output means for outputting a prompt generated based on a chat message received from the terminal to the target model when the chat message is received; a chat message output means for outputting to the terminal a chat message based on a response from the target model to the prompt; a search query output means for outputting a search query generated based on the log to a search system associated with the target model in response to a given condition being satisfied in a situation where the target model is a language model associated with a search engine; a search result output means for outputting search results received from the search system to the terminal in response to the search query being output to the search system; A search control system including:

2. A determination means for repeatedly executing a process of determining a target model from among a plurality of language models associated with search systems having different search targets, based on a log of chat message exchanges between the terminals; a prompt output means for outputting a prompt generated based on a chat message received from the terminal to the target model when the chat message is received; a chat message output means for outputting to the terminal a chat message based on a response from the target model to the prompt; a search query output means for outputting a search query generated based on the log to a search system associated with the target model in response to a given condition being satisfied; a search result output means for outputting to the terminal a search result for a search target of the search system, the search result being received from the search system in response to an output of the search query to the search system; A search control system including:

3. A determination means for repeatedly executing a process of determining a target model from among a plurality of language models based on a log of chat message exchanges between the terminal and the target model; a prompt output means for outputting a prompt generated based on a chat message received from the terminal to the target model when the chat message is received; a chat message output means for outputting to the terminal a chat message based on a response from the target model to the prompt; a search query output means for outputting a search query generated based on the log to a search system; a search result output means for outputting search results received from the search system to the terminal in response to output of the search query to the search system, all of the plurality of language models are able to refer to the log; Search control system.

4. A determination means for repeatedly executing a process of determining a target model from among a plurality of language models based on a log of chat message exchanges between the terminal and the target model; a prompt output means for outputting a prompt generated based on a chat message received from the terminal to the target model when the chat message is received; a chat message output means for outputting to the terminal a chat message based on a response from the target model to the prompt; a search query output means for outputting a search query generated based on the log to a search system; a search result output means for outputting search results received from the search system to the terminal in response to output of the search query to the search system, all of the plurality of language models are pre-configured to generate answers in a common communication style; Search control system.

5. the determination means executes a process of determining the target model every time a chat message is received from the terminal; The search control system according to any one of claims 1 to 4.

6. a search query generation means for generating a search query generation prompt that instructs generation of the search query based on the log, and for generating the search query based on a response output from the target model in response to output of the search query generation prompt to the target model, The search control system according to any one of claims 1 to 4.

7. The log includes the search results. The search control system according to any one of claims 1 to 4.

8. The search system executes a search targeting a product or service. The search control system according to any one of claims 1 to 4.

9. The search query is a keyword used to search for a product or service. The search control system according to any one of claims 1 to 4.

10. the search results indicate products or services; The search control system according to any one of claims 1 to 4.

11. The terminal further includes: The terminal includes a display means for displaying a screen on which chat messages and the search results are arranged in chronological order. The search control system according to any one of claims 1 to 4.

12. A determination means for repeatedly executing a process of determining a target template from a plurality of templates, each of which is either a template associated with the search system or a template not associated with the search system, based on a log of chat message exchanges between the terminal and the target template; a prompt output means for outputting a prompt generated based on a chat message received from the terminal and the target template when the chat message is received to a language model; a chat message output means for outputting to the terminal a chat message based on a response from the language model to the prompt; a search query output means for outputting a search query generated based on the log to the search system associated with the target template in response to a given condition being satisfied in a situation where the target template is a template associated with the search system; a search result output means for outputting search results received from the search system to the terminal in response to the search query being output to the search system; A search control system including:

13. A determination means for repeatedly executing a process of determining a target template from among a plurality of templates associated with search systems having different search targets, based on a log of chat message exchanges between the terminals; a prompt output means for outputting a prompt generated based on a chat message received from the terminal and the target template when the chat message is received to a language model; a chat message output means for outputting to the terminal a chat message based on a response from the language model to the prompt; a search query output means for outputting a search query generated based on the log to a search system associated with the target template in response to a given condition being satisfied; a search result output means for outputting to the terminal a search result for a search target of the search system, the search result being received from the search system in response to an output of the search query to the search system; A search control system including:

14. the determination means executes a process of determining the target template every time a chat message is accepted from the terminal.

14. The search control system according to claim 12 or 13.

15. The log includes the search results.

14. The search control system according to claim 12 or 13.

16. The search system executes a search targeting a product or service.

14. The search control system according to claim 12 or 13.

17. The search query is a keyword used to search for a product or service.

14. The search control system according to claim 12 or 13.

18. the search results indicate products or services; 14. The search control system according to claim 12 or 13.

19. The terminal further includes: The terminal includes a display means for displaying a screen on which chat messages and the search results are arranged in chronological order.

14. The search control system according to claim 12 or 13.

20. a prompt output means for outputting a prompt generated based on a chat message received from the terminal to the language model; a chat message output means for outputting to the terminal a chat message based on a response from the language model to the prompt; a search query output means for outputting a search query generated based on a log of chat message exchanges with the terminal to a search system; a search result output means for outputting search results received from the search system to the terminal in response to output of the search query to the search system, the chat message output means outputs a predetermined chat message indicating that a search is possible when a given condition is satisfied; the search query output means outputs the search query to the search system in response to a search request received in response to the output of the predetermined chat message; Search control system.

21. the search query output means outputs the search query to the search system in response to reception of a chat message indicating a request for a search for the output of the predetermined chat message; The search control system according to claim 20.

22. The log includes the search results. The search control system according to claim 20.

23. The search system executes a search targeting a product or service. The search control system according to claim 20.

24. The search query is a keyword used to search for a product or service. The search control system according to claim 20.

25. the search results indicate products or services; The search control system according to claim 20.

26. The terminal further includes: The terminal includes a display means for displaying a screen on which chat messages and the search results are arranged in chronological order. The search control system according to claim 20.

27. ​​A step in which a determination means repeatedly executes a process of determining a target model from a plurality of language models, each of which is either a language model associated with a search system or a language model not associated with a search system, based on a log of chat message exchanges between the terminals; a step of outputting a prompt generated based on a chat message received from the terminal to the target model when the chat message is received by a prompt output means; a chat message output means for outputting to the terminal a chat message based on a response from the target model to the prompt; a step in which a search query output means outputs a search query generated based on the log to a search system associated with the target model in response to a given condition being satisfied in a situation in which the target model is a language model associated with a search engine; a step of outputting search results received from the search system in response to the output of the search query to the search system by a search result output means to the terminal; A search control method including:

28. A step in which a determination means repeatedly executes a process of determining a target model from among a plurality of language models associated with search systems having different search targets, based on a log of chat message exchanges between the terminals; a step of outputting a prompt generated based on a chat message received from the terminal to the target model when the chat message is received by a prompt output means; a chat message output means for outputting to the terminal a chat message based on a response from the target model to the prompt; a search query output means outputting a search query generated based on the log to a search system associated with the target model in response to a given condition being satisfied; a step in which a search result output means outputs to the terminal a search result for a search target of the search system, the search result being received from the search system in response to an output of the search query to the search system; A search control method including:

29. A step in which a determination means repeatedly executes a process of determining a target model from among a plurality of language models based on a log of chat message exchanges with a terminal; a step of outputting a prompt generated based on a chat message received from the terminal to the target model when the chat message is received by a prompt output means; a chat message output means for outputting to the terminal a chat message based on a response from the target model to the prompt; a step in which a search query output means outputs a search query generated based on the log to a search system; a step in which a search result output means outputs search results received from the search system in response to output of the search query to the search system to the terminal; all of the plurality of language models are able to refer to the log; Search control methods.

30. A step in which a determination means repeatedly executes a process of determining a target model from among a plurality of language models based on a log of chat message exchanges with a terminal; a step of outputting a prompt generated based on a chat message received from the terminal to the target model when the chat message is received by a prompt output means; a chat message output means for outputting to the terminal a chat message based on a response from the target model to the prompt; a step in which a search query output means outputs a search query generated based on the log to a search system; a step in which a search result output means outputs search results received from the search system in response to output of the search query to the search system to the terminal; all of the plurality of language models are pre-configured to generate answers in a common communication style; Search control methods.

31. A step in which a determination means repeatedly executes a process of determining a target template from a plurality of templates, each of which is either a template associated with a search system or a template not associated with a search system, based on a log of chat message exchanges with a terminal; a step in which a prompt output means outputs a prompt generated based on a chat message received from the terminal and the target template when the chat message is received to a language model; a step of outputting a chat message based on a response from the language model to the terminal by a chat message output means; a step in which a search query output means outputs a search query generated based on the log to a search system associated with the target template in response to a given condition being satisfied in a situation in which the target template is a template associated with the search system; a step of outputting search results received from the search system in response to the output of the search query to the search system by a search result output means to the terminal; A search control method including:

32. A step in which a determination means repeatedly executes a process of determining a target template from a plurality of templates associated with search systems having different search targets based on a log of chat message exchanges with the terminal; a step in which a prompt output means outputs a prompt generated based on a chat message received from the terminal and the target template when the chat message is received to a language model; a step of outputting a chat message based on a response from the language model to the terminal by a chat message output means; a step of outputting a search query generated based on the log to a search system associated with the target template by a search query output means in response to a given condition being satisfied; a step in which a search result output means outputs to the terminal a search result for a search target of the search system, the search result being received from the search system in response to an output of the search query to the search system; A search control method including:

33. A step in which a prompt output means outputs a prompt generated based on a chat message received from a terminal to a language model; a step of outputting a chat message based on a response from the language model to the terminal by a chat message output means; a step in which a search query output means outputs a search query generated based on a log of chat message exchanges with the terminal to a search system; a step in which a search result output means outputs search results received from the search system in response to output of the search query to the search system to the terminal; In the step of outputting the chat message to the terminal, a predetermined chat message indicating that a search is possible is output in response to a given condition being satisfied; and in the step of outputting the search query to a search system, the search query is output to the search system in response to reception of a search request for the output of the predetermined chat message. Search control methods.

34. A step of repeatedly executing a process of determining a target model from a plurality of language models, each of which is either a language model associated with a search system or a language model not associated with a search system, based on a log of chat message exchanges between the terminals; outputting a prompt generated based on a chat message received from the terminal to the target model when the chat message is received; outputting to the terminal a chat message based on a response from the target model to the prompt; outputting a search query generated based on the log to a search system associated with the target model in response to a given condition being satisfied in a situation where the target model is a language model associated with a search engine; outputting search results received from the search system in response to the search query being output to the terminal; A program that causes a computer to execute the following.

35. A step of repeatedly executing a process of determining a target model from among a plurality of language models associated with search systems having different search targets, based on a log of chat message exchanges between terminals; outputting a prompt generated based on a chat message received from the terminal to the target model when the chat message is received; outputting to the terminal a chat message based on a response from the target model to the prompt; outputting a search query generated based on the log to a search system associated with the target model in response to a given condition being satisfied; outputting to the terminal a search result for a search target of the search system, the search result being received from the search system in response to an output of the search query to the search system; A program that causes a computer to execute the following.

36. A step of repeatedly executing a process of determining a target model from among a plurality of language models based on a log of chat message exchanges with a terminal; outputting a prompt generated based on a chat message received from the terminal to the target model when the chat message is received; outputting to the terminal a chat message based on a response from the target model to the prompt; outputting a search query generated based on the log to a search system; and outputting search results received from the search system to the terminal in response to output of the search query to the search system. all of the plurality of language models are able to refer to the log; program.

37. A step of repeatedly executing a process of determining a target model from among a plurality of language models based on a log of chat message exchanges with a terminal; outputting a prompt generated based on a chat message received from the terminal to the target model when the chat message is received; outputting to the terminal a chat message based on a response from the target model to the prompt; outputting a search query generated based on the log to a search system; and outputting search results received from the search system to the terminal in response to output of the search query to the search system. all of the plurality of language models are pre-configured to generate answers in a common communication style; program.

38. A step of repeatedly performing a process of determining a target template from a plurality of templates, each of which is either a template associated with a search system or a template not associated with a search system, based on a log of chat message exchanges between the terminals; outputting a prompt generated based on a chat message received from the terminal and the target template when the chat message is received to a language model; outputting to the terminal a chat message based on a response from the language model to the prompt; outputting a search query generated based on the log to a search system associated with the target template in response to a given condition being satisfied in a situation where the target template is a template associated with the search system; outputting search results received from the search system in response to the search query being output to the terminal; A program that causes a computer to execute the following.

39. A step of repeatedly executing a process of determining a target template from a plurality of templates associated with search systems having different search targets based on a log of chat message exchanges between the terminals; outputting a prompt generated based on a chat message received from the terminal and the target template when the chat message is received to a language model; outputting to the terminal a chat message based on a response from the language model to the prompt; outputting a search query generated based on the log to a search system associated with the target template in response to a given condition being satisfied; outputting to the terminal a search result for a search target of the search system, the search result being received from the search system in response to an output of the search query to the search system; A program that causes a computer to execute the following.

40. outputting prompts generated based on chat messages received from the terminal to a language model; outputting to the terminal a chat message based on a response from the language model to the prompt; outputting a search query generated based on a log of chat message exchanges with the terminal to a search system; outputting search results received from the search system to the terminal in response to output of the search query to the search system; In the step of outputting the chat message to the terminal, a predetermined chat message indicating that a search is possible is output in response to a given condition being satisfied; and in the step of outputting the search query to a search system, the search query is output to the search system in response to reception of a search request for the output of the predetermined chat message. program.

Citation Information

Patent Citations

  • Conversation generation method, conversation generation device and storage medium

    CN116226344A

  • Query method, device and system

    CN116701451A

  • Model calling method and device, electronic equipment and storage medium

    CN117973523A

  • Device, method and program for supporting commodity retrieval

    JP2011133926A

  • Using large language models in generating automated assistant responses

    JP2024521053A