Information processing system, information processing method, and program

Abstract

To improve the quality of answers from a large-scale language model to questions from questioners.SOLUTION: In a CPU 11 of a management server 1, a data acquisition unit 32 acquires an input query, and a language determination prompt generation unit 41 generates a language determination prompt for causing a large-scale language model to determine the language of the query. A language determination result acquisition unit 42 acquires a result of query language determination performed by the large-scale language model based on the language determination prompt. An answer generation prompt generation unit 51 generates an answer generation prompt for causing the large-scale language model to generate an answer to the query in the language of the query.SELECTED DRAWING: Figure 3

Description

[Technical Field]

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

[0002] There is a growing need for services that can support multiple languages ​​in a variety of services. For example, in web services, there is a demand not only for service content to be translated into multiple languages, but also for responses to user questions to be provided in multiple languages. However, for service providers, it is costly in terms of both money and time to hire personnel who can respond to multiple languages. In this situation, AI (artificial intelligence) technology has been developed in recent years that uses large language models (LLMs) to support multiple languages ​​and enable appropriate responses (e.g., Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2019-160286 Summary of the Invention [Problem to be solved by the invention]

[0004] In a system that answers questions using a predetermined language model, including a large-scale language model, it is possible to operate in a way that a human responder responds only when the answer from the language model is insufficient, thereby reducing the cost required to recruit human responders. On the other hand, there is also a demand for improvements in the accuracy of the answers, which indicates the quality of the language model's answers, and the accuracy of appropriately responding to the request of the questioner.

[0005] The present invention has been made in view of the above circumstances, and aims to improve the quality of answers provided by language models to questions from questioners. [Means for solving the problem]

[0006] In order to achieve the above object, one aspect of the present invention is to a query acquisition means for acquiring a query, which is the content of an input inquiry; a language determination processing means for executing a process of determining the language of the query using a language model; an answer generation processing means for executing a process of generating an answer to the query using the language model together with an instruction to generate an answer in the language determined as a result of the language determination; The information processing system includes:

[0007] An information processing method and a program corresponding to the information processing system according to one aspect of the present invention are also provided as an information processing method and a program corresponding to the information processing system according to one aspect of the present invention. [Effects of the Invention]

[0008] According to the present invention, the quality of answers from large-scale language models to questions from questioners is improved. [Brief explanation of the drawings]

[0009] [Figure 1] 1 is a diagram illustrating an example of the overall configuration of an information processing system according to an embodiment of the present invention. [Figure 2] FIG. 2 is a block diagram illustrating an example of a hardware configuration of a management server according to the present embodiment. [Figure 3] FIG. 2 is a functional block diagram showing an example of a functional configuration of a management server and a person in charge terminal. [Figure 4] 10 is a flowchart illustrating an example of a processing flow in a control unit of a management server. [Figure 5] FIG. 10 is a diagram showing a specific example of a user interface for inputting a query, which is displayed on a terminal of a person in charge. [Figure 6] FIG. 10 illustrates an example of a language determination prompt. [Figure 7]FIG. 10 illustrates an example of an answer generation prompt. DETAILED DESCRIPTION OF THE INVENTION

[0010] Hereinafter, this embodiment will be described with reference to the drawings. <Information Processing System> FIG. 1 is a diagram showing an example of the overall configuration of an information processing system S according to this embodiment. The information processing system S is a system that can be applied to the management operations of, for example, companies known as multinational corporations or global corporations whose customers and employees span the entire world. In this embodiment, an example will be described in which the information processing system S is applied to the management operations of a company whose employees each speak multiple languages. Hereinafter, the person in charge of the management operations to which the information processing system S is applied will be referred to as the "person in charge."

[0011] The information processing system S is configured by connecting a management server 1 and staff member terminals 2-1 to 2-n (n is an integer value of 1 or greater) via a network N. The network N is, for example, the Internet, a LAN (Local Area Network), or a VPN (Virtual Private Network). When there is no need to explain each of the staff member terminals 2-1 to 2-n individually, they will be collectively referred to as the "staff member terminal 2."

[0012] (Management Server) The management server 1 is an information processing device that serves as a server that manages the entire information processing system S. The management server 1 acquires various information transmitted from the person in charge terminal 2 and from the outside, and performs various processes. The management server 1 also transmits various information to the person in charge terminal 2 and the outside.

[0013] For example, the management server 1 acquires a query, which is the content of an inquiry input to the person in charge terminal 2. The management server 1 executes control to generate an answer to the query using a predetermined AI model, and provides the answer to the person in charge.

[0014] In the following, the present embodiment will be described assuming that a generative AI model employing a large-scale language model (LLM) is used as the predetermined AI model. A generative AI model is an AI model capable of outputting new data and information based on data learned by a computer. Furthermore, when a prompt (character string) using natural language is input, the language model can output the character string based on the input prompt and the context of prompts previously input and output. In particular, a large-scale language model improves the accuracy of the answer, which indicates the quality of the answer, and the accuracy of appropriately responding to the request of the questioner, by increasing the scale of elements of the language model, such as the "computational volume," "data volume," and "number of model parameters." As described above, the present embodiment will be described assuming that a large-scale language model is employed. However, the present embodiment is not limited to this, and a language model using a predetermined natural language may also be employed. In the description of this embodiment, the expression "taking into account predetermined matters in the processing of a language model" means executing processing to generate an output from the language model based on the predetermined matters. Specifically, for example, this is realized by entering an instruction sentence or premise information in a prompt of the language model so that an answer is generated based on the predetermined matters, and then output that reflects the instruction and premise knowledge.

[0015] In processing using a large-scale language model, prompt engineering is performed to obtain more suitable output strings and to specify the output format. In other words, rather than simply using questions, requests, etc. as prompts, prompt engineering embeds information such as assumptions, constraints, and output format specifications, resulting in more appropriate content for the output string.

[0016] In this embodiment, as described above, a plurality of prepared prompt templates are used to generate various types of prompts (including the generation of determination results) for a general-purpose large-scale language model in order to provide more suitable outputs.

[0017] In the following description of this embodiment, it is assumed that prompt templates are prepared in advance for each processing content (for example, language determination, query-based answer generation, knowledge information extraction, etc., as described below), but this is not particularly limited to this. That is, for example, the parts that are referred to as templates in the description of this embodiment may be generated or adjusted as appropriate.

[0018] The management server 1 pre-stores, as a language determination prompt template, a character string for determining the language of the query using a large-scale language model. That is, the language determination prompt template is a character string created in a predetermined language by prompt engineering in order to improve the accuracy of determining the language of the query and obtain a language determination result in a predetermined format. In the following description of this embodiment, the language determination prompt template is assumed to be an English character string.

[0019] The management server 1 generates a prompt for language determination (appropriately referred to as a "language determination prompt") by adding a query or the like to this language determination prompt template.

[0020] When generating a language determination prompt, the management server 1 can generate a language determination prompt that includes taking into account the standard time zone of the user who input the query in determining the language of the query using a large-scale language model.Furthermore, when generating a language determination prompt, the management server 1 can generate a language determination prompt that includes taking into account the context of the query in determining the language of the query using a large-scale language model.

[0021] The management server 1 acquires the result of the language determination of the query performed by the large-scale language model based on the generated language determination prompt. Then, the management server 1 generates a character string (hereinafter referred to as an "answer generation prompt") in a predetermined language to cause the large-scale language model to generate an answer to the query in the language of the query. In this embodiment, the explanation will be given assuming that the language of the answer generation prompt is English.

[0022] When generating an answer generation prompt, the management server 1 can generate an answer generation prompt that includes knowledge related to the query extracted from information (hereinafter referred to as "knowledge") stored in a predetermined database, by using various types of information related to the management work of the person in charge as knowledge that contributes to answering the query. The specific configuration and processing details of the management server 1 will be described later.

[0023] (Personnel terminal) The staff member terminal 2 is an information processing device operated by the staff member, and is configured, for example, as a personal computer, a tablet terminal, a smartphone, etc. It is assumed that dedicated application software that enables the information processing system S to be used is installed on the staff member terminal 2.

[0024] The person in charge terminal 2 starts dedicated application software based on the operation of the person in charge, and performs various processes based on various information transmitted from the management server 1 and external sources, information input by the operation of the person in charge, etc. The person in charge terminal 2 also transmits various information to the management server 1 and external sources.

[0025] For example, the person in charge terminal 2 accepts an operation to input a query and transmits the input information to the management server 1. Also, for example, the person in charge terminal 2 acquires a response to the query transmitted from the management server 1 and displays it on a display or the like. The specific configuration and processing details of the person in charge terminal 2 will be described later.

[0026] The above-described processes performed by the management server 1 and the person in charge terminal 2 that constitute the information processing system S are merely examples, and it is sufficient that the information processing system S as a whole has the functions to realize the above-described processes. Therefore, some or all of the functions to realize the above-described processes may be shared or cooperated within the information processing system S.

[0027] For example, some or all of the functions of the management server 1 may be functions of other information processing devices, etc. in the information processing system S, and some or all of the functions of other information processing devices, etc. in the information processing system S may be functions of the management server 1. Furthermore, some or all of the functions of the management server 1 may be transferred to other servers, etc. (not shown). This promotes processing in the information processing system S as a whole and also makes it possible for processes to complement each other.

[0028] <Hardware configuration> (Management server hardware configuration) FIG. 2 is a block diagram showing an example of the hardware configuration of the management server 1 according to this embodiment. The management server 1 includes a CPU (Central Processing Unit) 11, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, a bus 14, an input / output interface 15, an output unit 16, an input unit 17, a memory unit 18, a communication unit 19, and a drive 20.

[0029] The CPU 11 executes various processes in accordance with programs recorded in the ROM 12 or programs loaded from the storage unit 18 into the RAM 13. The RAM 13 also stores data and the like required for the CPU 11 to execute various processes. The CPU 11, the ROM 12, and the RAM 13 are interconnected via a bus 14. An input / output interface 15 is also connected to this bus 14.

[0030] The input / output interface 15 is connected to an output unit 16, an input unit 17, a memory unit 18, a communication unit 19, and a drive 20. The output unit 16 is composed of a display, a speaker, etc., and outputs various types of information as images and sounds. The input unit 17 is composed of a keyboard, a mouse, a touch panel, etc., and accepts input of various types of information. The memory unit 18 is composed of a hard disk, a DRAM (Dynamic Random Access Memory), etc., and stores various types of data. The communication unit 19 communicates with other devices via a network N including the Internet.

[0031] Removable media 21, such as a magnetic disk, optical disk, magneto-optical disk, or semiconductor memory, is appropriately attached to the drive 20. Programs read from the removable media 21 by the drive 20 are installed in the storage unit 18 as needed. The removable media 21 can also store various data stored in the storage unit 18 in the same way as the storage unit 18.

[0032] (Hardware configuration of the staff terminal) Although not shown, the person in charge terminal 2 has the hardware configuration shown in Fig. 2. That is, the person in charge terminal 2 has a CPU, ROM, RAM, bus, input / output interface, output unit, input unit, memory unit, communication unit, drive, and removable media that correspond to the CPU 11, ROM 12, RAM 13, bus 14, input / output interface 15, output unit 16, input unit 17, memory unit 18, communication unit 19, drive 20, and removable media 21 of the management server 1 in Fig. 2, respectively.

[0033] <Functional configuration> (Functional configuration of management server and staff terminal) 3 and 4 are functional block diagrams showing an example of the functional configuration of the management server 1 and the person in charge terminal 2. As shown in FIG. In operation, the CPU 11 of the management server 1 functions as a data management unit 31, a data acquisition unit 32, a language determination processing unit 33, an answer generation processing unit 34, a large-scale language model processing unit 35, and a transmission control unit 36. The language determination processing unit 33 has a language determination prompt generation unit 41 and a language determination result acquisition unit 42. The answer generation processing unit 34 has an answer generation prompt generation unit 51 and an answer generation result acquisition unit 52. The storage unit 18 of the management server 1 is provided with a query DB 61 in which acquired queries are stored, a template DB 62 in which generated language determination prompts and answer generation prompts are stored, and a knowledge DB 63 in which knowledge is stored in advance.

[0034] The data management unit 31 manages various types of data. For example, the data management unit 31 stores and manages acquired queries in a query DB 61. The data management unit 31 also stores and manages generated language determination prompts and answer generation prompts in a template DB 62. The data management unit 31 also stores and manages knowledge in a knowledge DB 63. The knowledge includes various types of information related to management tasks, such as information about various types of employment and salaries in various types of countries.

[0035] The data acquisition unit 32 acquires various types of information. For example, the data acquisition unit 32 acquires a query that has been input to and transmitted from the person in charge terminal 2. The data acquisition unit 32 also acquires knowledge related to the query that has been extracted from the knowledge DB 63 in the storage unit 18.

[0036] The language determination processing unit 33 executes a process for determining the language of a query.

[0037] Specifically, for example, the language determination prompt generation unit 41 of the language determination processing unit 33 generates a language determination prompt template, which is a string of characters for determining the language of a query using a large-scale language model, and a language determination prompt to be input to the large-scale language model based on the query. A specific example of the language determination prompt and its details will be described later with reference to FIG. 6, but the language determination prompt is made up of a template portion in a predetermined language (for example, English) and a query in the language to be determined.

[0038] The language determination prompt generation unit 41 generates a language determination prompt that includes taking into account a standard time zone in determining the language of a query using a large-scale language model when generating the language determination prompt. Also, the language determination prompt generation unit 41 generates a language determination prompt that includes taking into account a context of the query in determining the language of the query using a large-scale language model when generating the language determination prompt.

[0039] Next, the language determination result acquisition unit 42 acquires the language determination result of the query using the large-scale language model based on the language determination prompt. Specifically, for example, the language determination result acquisition unit 42 executes processing using a large-scale language model based on the language determination prompt generated by the language determination prompt generation unit 41, and acquires the output result as the language determination result.

[0040] In this way, the language determination processing unit 33 inputs the language determination prompt generated by the language determination prompt generation unit 41 into the large-scale language model, and performs processing to treat the output from the large-scale language model as the language determination result. In this embodiment, the large-scale language model processing unit 35 provided in the management server 1 (described later) can execute the processing using the large-scale language model, but the present invention is not limited to this. For example, the large-scale language model processing may be executed by another information processing device (not shown).

[0041] The answer generation processing unit 34 executes processing to generate an answer to a query.

[0042] Specifically, for example, first, the answer generation prompt generating unit 51 of the answer generation processing unit 34 generates an answer generation prompt for generating an answer to a query in the language of the query using a large-scale language model. A specific example of an answer generation prompt and its details will be described later with reference to FIG. 7. The answer generation prompt is made up of a template portion in a predetermined language (for example, English) and the query language determination result.

[0043] When generating an answer generation prompt, the answer generation prompt generation unit 51 can generate an answer generation prompt that includes knowledge related to the query extracted from knowledge stored in the knowledge DB 63. The extraction of knowledge related to the query from the knowledge DB 63 is performed by a large-scale language model. When causing the large-scale language model to extract knowledge, the result of determining the language of the query may be used, or the result of determining the language of the query may not be used. Furthermore, when causing the large-scale language model to extract knowledge, information regarding the country to which the question is addressed may be taken into consideration.

[0044] Next, the answer generation result acquisition unit 52 acquires an answer generation result for the query using the large-scale language model based on the answer generation prompt. Specifically, for example, the answer generation result acquisition unit 52 executes processing using a large-scale language model based on the answer generation prompt generated by the answer generation prompt generation unit 51, and acquires the output result as the language determination result.

[0045] In this way, the answer generation processing unit 34 inputs the answer generation prompt generated by the answer generation prompt generation unit 51 into the large-scale language model, and performs processing to determine the result of language determination based on the output from the large-scale language model. In this embodiment, the large-scale language model processing unit 35 provided in the management server 1 (described later) can execute the processing using the large-scale language model, but the present invention is not limited to this. For example, the large-scale language model processing may be executed by another information processing device (not shown).

[0046] When a prompt (character string) using a natural language is input, the large-scale language model processing unit 35 uses a large-scale language model stored in the large-scale language model DB 64 to execute processing for outputting the character string.

[0047] The transmission control unit 36 ​​controls the transmission of various information to the person in charge terminal 2 and to the outside via the communication unit 19. For example, the transmission control unit 36 ​​controls the transmission of a response to a query to the person in charge terminal 2.

[0048] In addition, in the CPU of the person in charge terminal 2, an acquisition unit 71, a display control unit 72, and a transmission control unit 73 function when the terminal 2 is operating.

[0049] The acquisition unit 71 acquires various types of information. For example, the acquisition unit 71 acquires input information accepted by the input unit. For example, the acquisition unit 71 acquires an input query. The acquisition unit 71 also acquires various types of information transmitted from the management server 1 and from outside. For example, the acquisition unit 71 acquires a response to a query transmitted from the management server 1.

[0050] The display control unit 72 controls the display of various types of information on the display of the output unit, etc. For example, the display control unit 72 controls the display of the output unit, etc., of a user interface for inputting a query. Furthermore, for example, the display control unit 72 controls the display of the output unit, etc., of an answer to the query acquired by the acquisition unit 71.

[0051] The transmission control unit 73 controls the transmission of various information to the management server 1 and to the outside via the communication unit. For example, the transmission control unit 73 controls the transmission of an input query to the management server 1.

[0052] <Processing flow> (Management Server Processing Flow) FIG. 4 is a flowchart showing an example of the flow of processing in the CPU 11 of the management server 1. In the processing in the CPU 11 of the management server 1, the data management unit 31 of the CPU 11 manages the knowledge stored in the knowledge DB 63 (step S1).

[0053] Thereafter, when a query is transmitted from the person in charge terminal 2 (YES in step S2), the data acquisition unit 32 of the CPU 11 acquires the transmitted query (step S3). Then, the data management unit 31 stores and manages the query acquired by the data acquisition unit 32 in a database (query DB 61 in FIG. 3) (step S4). On the other hand, when a query has not been transmitted (NO in step S2), the data acquisition unit 32 repeats the determination process of step S2 until a query is transmitted.

[0054] Next, the language determination prompt generation unit 41 of the CPU 11 generates a language determination prompt for determining the language of the query acquired in step S3 (step S5). The generated language determination prompt may include a prompt to take into account the standard time zone when determining the language of the query using a large-scale language model.

[0055] Next, the language determination result acquisition unit 42 of the CPU 11 inputs the language determination prompt generated in step S5 into the large-scale language model (step S6), and determines the language of the query by obtaining the language determination result output from the large-scale language model (step S7).

[0056] Next, the answer generation prompt generation unit 51 of the CPU 11 generates an answer generation prompt for causing the large-scale language model to generate an answer to the query in the language of the query (step S8). The generated answer generation prompt may include knowledge related to the query extracted by the large-scale language model from knowledge stored in the knowledge DB 63.

[0057] Next, the answer generation result acquisition unit 52 of the CPU 11 inputs the answer generation prompt generated in step S8 into the large-scale language model (step S9), and generates an answer to the query by acquiring the answer generation result output from the large-scale language model (step S10). Then, the transmission control unit 36 ​​of the CPU 11 controls transmission of the answer to the query to the person in charge terminal 2 (step S11). This ends the processing in the CPU 11 of the management server 1 (END).

[0058] <Example> (Specific examples of various types of data) FIG. 5 is a diagram showing a specific example of a user interface displayed on the person in charge terminal 2 for inputting a query. Display area F1 of the user interface shown in Figure 5 displays a text message M1 that reads, "AI will answer any questions you have about employment policies around the world. Please enter the country you are asking about and the content of your question." Also displayed in display area F1 are a button B1 for displaying a pull-down menu for selecting the country that the person in charge will be asked about (hereinafter referred to as the "question country"), an input field R1 for entering a query (question), and a button B2 labeled "ANSWER."

[0059] For example, let's assume that a person whose native language is Japanese wants to inquire about the probationary period of an employee in Argentina. In this case, the person presses button B1 to display a pull-down menu for selecting the country of the question, and selects "Argentina." The person then enters the query "Please tell me about the probationary period" in Japanese in input field R1, and presses button B2 to display the answer.

[0060] Then, the answer to the query is displayed in Japanese in the display area F2 of the user interface, as shown in Fig. 5. That is, the information processing system S determines that the language of the query is Japanese based on the content and context of the query entered in the input field R1, and displays the answer to the query in Japanese.

[0061] Also, although not shown, suppose that a person whose native language is English enters a query in English, "Please tell me about the trial period," in input field R1 to inquire about the probationary period of an employee in Argentina, and presses button B2 to display the answer. In this case, the answer to the query is displayed in English in display area F2 of the user interface. That is, information processing system S determines that the language of the query is English based on the content, context, etc. of the query entered in input field R1, and displays the answer to the query in English.

[0062] The same applies even if the agent's native language is a language other than Japanese or English. If the query language is determined to be Chinese based on the content and context of the query entered in input field R1, the query response will be displayed in Chinese, and if the query language is determined to be Korean based on the content and context of the query entered in input field R1, the query response will be displayed in Korean.

[0063] FIG. 6 shows a specific example of a language determination prompt. The language determination prompt shown in Figure 6 is generated in English. The language determination prompt is character string data used to improve the accuracy of the answer (query language determination result) from the large-scale language model. The language determination prompt includes information such as the position (overview) of the large-scale language model, "GOALS" indicating the goal of the answer, "Context" indicating the context that is the premise for the answer, "Constraints" indicating the constraints on the answer, and "Performance Evaluation" indicating the evaluation criteria for verifying the accuracy of the answer.

[0064] Among the strings included in the language determination prompt, "Context" includes "timezone," which indicates the standard time zone of the person in charge. For example, if the person in charge is in Japan, "Japan Standard Time, JST" is entered as "timezone." "Timezone" is automatically obtained, for example, from the setting information of the person in charge's terminal 2 operated by the person in charge, and environmental information such as the web browser and application software running on the person in charge's terminal 2.

[0065] FIG. 7 shows an example of an answer generation prompt. The answer generation prompt shown in Figure 7 is generated in English. The answer generation prompt is character string data used to improve the accuracy of the answer (query answer) from the large-scale language model. Like the language determination prompt described above, the answer generation prompt includes the position (overview) of the large-scale language model, "GOALS" indicating the goal of the answer, "Context" indicating the context that is the premise for the answer, "Constraints" indicating the constraints on the answer, and "Performance Evaluation" indicating the evaluation criteria for verifying the accuracy of the answer.

[0066] Among the strings included in the answer generation prompt, "Constraints" includes "countries" which indicates the country of the question and "language" which indicates the language that has been determined. Also, among the strings included in the answer generation prompt, "Context" includes "document" which indicates knowledge (an example of knowledge information).

[0067] <Advantageous Effects of the Present Embodiment> According to the above-described embodiment, a language determination prompt in which the agent's standard time zone and a query are embedded is input to a large-scale language model, and a query language determination result is output. Then, an answer generation prompt in which the query, the query language determination result, the question target country, and knowledge are embedded is input to a large-scale language model, and an answer to the query is output. As a result, the answer is displayed in the language of the query entered by the agent into the user interface. For example, even if the language of the query differs from the language of the language determination prompt or the answer generation prompt, the answer is output in the query language without being affected by the language of the language determination prompt or the answer generation prompt. As a result, the quality of the answers provided by the large-scale language model to questions from questioners is improved.

[0068] <Other embodiments> Although one embodiment of the present invention has been described above, the present invention is not limited to the above-described embodiment, and modifications, improvements, etc. within the scope of achieving the object of the present invention are included in the present invention.

[0069] For example, the above-described series of processes can be executed by hardware or software. In other words, the above-described functional configuration is merely an example and is not particularly limited. In other words, it is sufficient for the information processing system to be provided with a function that can execute the above-described series of processes as a whole, and the type of functional block used to realize this function is not particularly limited to the above-described example.

[0070] The location of the functional blocks is not particularly limited and may be arbitrary. For example, the functional blocks of the management server 1 may be transferred to another device, or the functional blocks of another device may be transferred to a server. Furthermore, one functional block may be configured as a single piece of hardware, a single piece of software, or a combination of both.

[0071] When a series of processes is executed by software, the programs constituting the software are installed onto a computer or the like from a network or a recording medium. The computer may be a computer incorporated into dedicated hardware. The computer may also be a computer capable of executing various functions by installing various programs, such as a server, a general-purpose smartphone, or a personal computer.

[0072] The recording medium containing such a program may be configured as a removable medium (not shown) that is distributed separately from the device main body in order to provide the program to users, etc., or may be configured as a recording medium that is pre-installed in the device main body and provided to users, etc. Since the program can be distributed via a network, the recording medium may be installed in or accessible from a computer that is connected or connectable to the network.

[0073] In this specification, the steps describing the program recorded on the recording medium include not only processes that are performed in chronological order, but also processes that are not necessarily performed in chronological order but are performed in parallel or individually. Also, in this specification, the term "system" means an overall device composed of multiple devices or multiple means, etc.

[0074] In other words, the information processing system to which the present invention is applied can take various forms having the following configurations. (1) That is, an information processing system to which the present invention is applied (for example, the information processing system S in FIG. 1) is a query acquisition unit (for example, the data acquisition unit 32 in FIG. 3) for acquiring a query, which is the content of an input inquiry; a language determination processing unit (for example, the language determination processing unit 33 in FIG. 3) that executes a process of determining the language of the query using a large-scale language model; an answer generation processing means (for example, the answer generation processing unit 34 in FIG. 3) that uses the large-scale language model to execute a process of answering the query together with an instruction to provide an answer in the language determined as a result of the language determination; It is enough to have this.

[0075] (2) The answer generation processing means an answer generation prompt generation unit (e.g., the answer generation prompt generation unit 51 in FIG. 3) that generates an answer generation prompt to be input to the large-scale language model based on an answer generation prompt template, which is a character string for causing the large-scale language model to provide an answer to the query, the result of the language determination, and the query; an answer generation result acquisition unit (e.g., the answer generation result acquisition unit 52 in FIG. 3) for acquiring the answer to the query generated by the large-scale language model based on the answer generation prompt; It can have or be.

[0076] (3) The language determination processing means a language determination prompt generation unit (e.g., a language determination prompt generation unit 41) that generates a language determination prompt to be input to the large-scale language model based on a language determination prompt template, which is a character string for determining the language of the query using a large-scale language model, and the query; a language determination result acquisition unit (e.g., the language determination result acquisition unit 42 in FIG. 3) for acquiring a result of the language determination of the query performed by the large-scale language model based on the language determination prompt; It can have or be.

[0077] (4) knowledge information acquisition means (for example, the data acquisition unit 32 in FIG. 3) for acquiring knowledge information related to the query based on the query; Furthermore, The answer generation prompt generating means and generating the answer generation prompt based on the knowledge information. 3. The information processing system according to claim 2.

[0078] (5) The knowledge information acquisition means extracting, as the knowledge information, knowledge related to the query from a database in which the knowledge information candidates are stored, using the large-scale language model; 5. The information processing system according to claim 4.

[0079] (6) The language determination prompt generation means generating the language determination prompt, the language determination prompt including taking into account the standard time zone of an information processing device that inputs the query; 4. The information processing system according to claim 3.

[0080] (7) The language determination prompt generation means generating the language determination prompt, the language determination prompt including considering a context of the query in determining the language of the query; 4. The information processing system according to claim 3.

[0081] (8) An information processing method executed by an information processing system, a query acquisition step of acquiring a query, which is the content of an input inquiry; a language determination processing step of performing a process of determining the language of the query using a large-scale language model; an answer generation processing step of executing a process to generate an answer to the query using the large-scale language model together with an instruction to generate an answer in the language determined as a result of the language determination; The information processing method may include the steps of:

[0082] (9) To the computer, a query acquisition step of acquiring a query, which is the content of an input inquiry; a language determination processing step of performing a process of determining the language of the query using a large-scale language model; an answer generation processing step of executing a process to generate an answer to the query using the large-scale language model together with an instruction to generate an answer in the language determined as a result of the language determination; The program may be a program that executes a control process including the above. [Explanation of symbols]

[0083] 1: Management server, 2: Person in charge terminal, 11: CPU, 16: Output unit, 17: Input unit, 18: Memory unit, 19: Communication unit, 31: Data management unit, 32: Data acquisition unit, 33: Language determination processing unit, 34: Answer generation processing unit, 35: Large-scale language model processing unit, 36: Transmission control unit, 41: Language determination prompt generation unit, 42: Language determination result acquisition unit, 51: Answer generation prompt generation unit, 52: Answer generation result acquisition unit, 71: Acquisition unit, 72: Display control unit, 73: Transmission control unit, S: Information processing system, N: Network

Claims

1. a query acquisition means for acquiring a query, which is the content of an input inquiry; a language determination processing means for executing a process of determining the language of the query using a large-scale language model; an answer generation processing means for executing a process of generating an answer to the query using the large-scale language model together with an instruction to generate an answer in the language determined as a result of the language determination; An information processing system comprising:

2. The answer generation processing means an answer generation prompt generation means for generating an answer generation prompt to be input to the large-scale language model based on an answer generation prompt template, which is a character string for causing the large-scale language model to provide an answer to the query, the result of the language determination, and the query; an answer generation result obtaining means for obtaining the answer to the query made by the large-scale language model based on the answer generation prompt; having The information processing system according to claim 1 .

3. The language determination processing means a language determination prompt generation means for generating a language determination prompt to be input to the large-scale language model based on a language determination prompt template, which is a character string for determining the language of the query using a large-scale language model, and the query; a language determination result obtaining means for obtaining a result of the language determination of the query performed by the large-scale language model based on the language determination prompt; having The information processing system according to claim 2 .

4. knowledge information acquiring means for acquiring knowledge information relating to the query based on the query; Furthermore, The answer generation prompt generating means and generating the answer generation prompt based on the knowledge information. The information processing system according to claim 2 .

5. The knowledge information acquisition means extracting, as the knowledge information, knowledge related to the query from a database in which the knowledge information candidates are stored, using the large-scale language model; The information processing system according to claim 4 .

6. The language determination prompt generation means generating the language determination prompt, the language determination prompt including taking into account the standard time zone of an information processing device that inputs the query; The information processing system according to claim 3 .

7. The language determination prompt generation means generating the language determination prompt, the language determination prompt including considering a context of the query in determining the language of the query; The information processing system according to claim 3 .

8. An information processing method executed by an information processing system, a query acquisition step of acquiring a query, which is the content of an input inquiry; a language determination processing step of performing a process of determining the language of the query using a large-scale language model; an answer generation processing step of executing a process to generate an answer to the query using the large-scale language model together with an instruction to generate an answer in the language determined as a result of the language determination; An information processing method including:

9. On the computer, a query acquisition step of acquiring a query, which is the content of an input inquiry; a language determination processing step of performing a process of determining the language of the query using a large-scale language model; an answer generation processing step of executing a process to generate an answer to the query using the large-scale language model together with an instruction to generate an answer in the language determined as a result of the language determination; A program that executes control processing including:

Images