Information processing apparatus, control method of information processing apparatus, and program

Abstract

To provide an information processing apparatus that can perform processing using a trained model including a large-scale language model without requiring a provider of target data to pay attention, and efficiently utilizes output data, a control method of the information processing apparatus, and a program.SOLUTION: An information processing apparatus 1 includes: a data receiving unit 30 which executes receiving processing to receive target data to be processed; an integrated processing unit 31 which executes acquisition processing to output a script including the target data and a command to a generative AI server 101 and obtain output data output by the generative AI server 101; and a data management unit 33 that can execute storage processing to store the output data in a database 90 and presentation processing to present the output data stored by the storage processing to an external terminal, such as a data user terminal 3.SELECTED DRAWING: Figure 7

Description

[Technical field]

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

[0002] Conventionally, there is known a technique for automatically responding to text data such as an inquiry input from a user terminal by a computer. For example, Patent Document 1 describes this type of technique.

[0003] Patent Document 1 relates to a question answering system including a user terminal device that outputs answers to questions input by users, and an answerer terminal device that transmits answers by answerers to the user terminal device. In Patent Document 1, the user terminal device outputs an automatic answer, which is an answer to a question that is automatically generated based on a learning result by machine learning, as an answer, and executes a process of connecting to the answerer terminal device when it is determined that the automatic answer does not reach an appropriate answer based on an evaluation index related to the answer to the question. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent Publication No. 2022-86817 Summary of the Invention [Problem to be solved by the invention]

[0005] In a system using a trained model including a large-scale language model, the answer output varies depending on the prompt settings. A user who provides target data to be processed is not necessarily aware of a trained model including a large-scale language model, and the target data is not necessarily input in a format suitable for a trained model including a large-scale language model. In addition, there are cases where it is desired to analyze meeting minutes, etc. as target data instead of questions, but even if the minutes are directly input into a trained model including a large-scale language model, the output data of the trained model including a large-scale language model is not necessarily in the desired format.

[0006] In addition, even if the output data output by a trained model including a large-scale language model is appropriate, the output data is returned to the person who input the target data. Therefore, although the person who provides the target data can check the output data, other data users who want to use the output result of the trained model including a large-scale language model cannot directly check it. Although it is possible for a data user to check the output data of the generated AI via the person who provides the target data, the conventional technology has room for improvement in terms of efficiently utilizing the output data of the trained model including a large-scale language model.

[0007] The present invention has been made in consideration of the above-mentioned circumstances, and aims to provide an information processing device, a control method for an information processing device, and a program that can perform processing using a trained model including a large-scale language model without the knowledge of the provider of the target data, and can efficiently utilize output data. [Means for solving the problem]

[0008] In order to achieve the above-mentioned object, one aspect of the present invention is an information processing device comprising: a data reception unit that executes a reception process to receive target data to be processed; a collaboration processing unit that executes an acquisition process to output a script including the target data and a command to a trained model including a large-scale language model and acquire output data output by the trained model; and a data management unit that executes a storage process to store the output data in a database and executes a presentation process to present the output data stored by the storage process to an external terminal. Effect of the Invention

[0009] According to the information processing device, the control method for the information processing device, and the program of the present invention, processing can be performed using a trained model including a large-scale language model without the provider of the target data being aware of it, and the output data can be utilized more efficiently. [Brief description of the drawings]

[0010] [Figure 1] 1 is a diagram showing a data processing system realized by an information processing device according to a first embodiment of the present invention. [Diagram 2] 1 is a block diagram showing a hardware configuration of an information processing device according to a first embodiment. [Diagram 3] 1 is a functional block diagram showing an example of a functional configuration of an information processing device according to a first embodiment. [Figure 4] 4 is a flowchart showing the flow of a determination process by the information processing device according to the first embodiment. FIG. [Diagram 5] FIG. 11 is a diagram showing an example of target data input from a target terminal. [Figure 6] FIG. 11 is a diagram illustrating an example of output data acquired by a collaboration processing unit. [Figure 7] 4 is a flowchart showing a flow of a storage process and a presentation process of output data by the information processing device according to the first embodiment. FIG. [Figure 8] FIG. 13 is a diagram showing an example of conversation content in a chat room for attendance reports as target data. [Figure 9] FIG. 13 is a diagram showing an example of output data after formatting created based on the conversation content in a chat room. [Figure 10] FIG. 13 is a diagram illustrating a state in which the contents of conversations between chat room members over a week as target data are registered in a database. [Figure 11] FIG. 13 is a diagram showing a schematic diagram in which summaries of conversation contents between members over the course of a week are registered in a database, enabling each member to check the summaries of conversation contents over the course of a week. [Figure 12] FIG. 11 is a functional block diagram showing an example of a functional configuration of an information processing device according to a second embodiment. [Figure 13] FIG. 11 is a flowchart showing a flow of extraction processing and linkage processing by an information processing device according to a second embodiment. [Figure 14] FIG. 11 is a flowchart showing the flow of a plurality of cooperative processes performed by an information processing device according to a third embodiment. [Figure 15] FIG. 13 is a functional block diagram showing an example of a functional configuration of an information processing device according to a fourth embodiment. [Figure 16] 13 is a flowchart showing an example of the flow of a process for creating a work procedure manual by an information processing device according to the fourth embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. Note that the embodiment described below is merely an example, and the technical scope of the present invention is not limited to this. Various modified examples are also included in the technical scope of the present invention.

[0012] First Embodiment FIG. 1 is a diagram showing a data processing system 100 realized by an information processing device 1 according to the first embodiment of the present invention.

[0013] <System configuration> First, the overall system configuration will be described. The data processing system 100 cooperates with an external generation AI (Artificial Intelligence) server 101 and SaaS (Software as a Service) 102 to process target data and scripts including commands input from a target terminal 2 or the like, based on a pre-set work procedure manual. The generation AI server 101 is a learned model including a large-scale language model that performs data processing on input target data such as text, audio, images, and videos, and scripts including commands, using a learning model constructed using techniques such as machine learning and deep learning.

[0014] The information processing device 1 is a server configured by a computer. The information processing device 1 transmits and receives various information to and from the target person terminal 2 and the data user terminal 3 via a communication network such as the Internet. The information processing device 1 may transmit and receive data to and from the target person terminal 2 and the data user terminal 3 via SaaS102.

[0015] The information processing device 1 processes target data in cooperation with the generation AI server 101 based on a preset work procedure manual. The work procedure manual specifies the contents and procedures for processing the target data set by a data user or the like.

[0016] The subject terminal 2 is used by a user who provides the target data to the information processing device 1. The subject terminal 2 is configured by a smartphone, a tablet, a personal computer, or the like. The subject terminal 2 may exchange various information with the information processing device 1 by a pre-installed program, may exchange various information through a web browser, or may exchange information using an email address. Note that FIG. 1 illustrates one of the multiple subject terminals 2.

[0017] The data user terminal 3 is used by a data user who uses the data processed by the information processing device 1. The data user terminal 3 is configured by a personal computer, a tablet, a smartphone, or the like. The data user terminal 3 may exchange various information with the information processing device 1 by a pre-installed program, may exchange various information through a web browser, or may exchange information using an email address. Note that FIG. 1 illustrates one of the multiple data user terminals 3. The data user terminal 3 may also be the same terminal as the target user terminal 2.

[0018] <Hardware configuration> Next, a description will be given of an example of hardware constituting the information processing device 1. Fig. 2 is a block diagram showing a hardware configuration of the information processing device 1 according to the first embodiment. The information processing device 1 includes a CPU (Central Processing Unit) 11 as a processor, 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 storage unit 18, a communication unit 19, and a drive 20.

[0019] The CPU 11 executes various processes according to a program recorded in the ROM 12 or a program loaded from the storage unit 18 to 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.

[0020] An output unit 16, an input unit 17, a storage unit 18, a communication unit 19, and a drive 20 are connected to the input / output interface 15. The output unit 16 is composed of a display, a speaker, etc., and outputs various information as images and sounds. The input unit 17 is composed of a keyboard, a mouse, etc., and inputs various information. The storage unit 18 is composed of a hard disk, a DRAM (Dynamic Random Access Memory), a SSD (Solid State Drive), etc., and stores various data. The communication unit 19 communicates with other devices via a network including the Internet.

[0021] Removable media 21, which may be a magnetic disk, an optical disk, a magneto-optical disk, or a 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 necessary. The removable media 21 can also store various data stored in the storage unit 18 in the same manner as the storage unit 18.

[0022] The hardware configuration described here is merely an example. The computer described in this embodiment including the information processing device 1 may have a configuration common to that of FIG. 2, or may have a different configuration. The computer may be composed of two or more computers. The subject terminal 2 and the data user terminal 3 in FIG. 1 are, for example, a smartphone, tablet, or personal computer having the same hardware configuration as that shown in FIG. 2. The computer may be realized by cloud computing using a communication network.

[0023] <Functional configuration> Next, a description will be given of the functional configuration of the information processing device 1. Fig. 3 is a functional block diagram showing an example of the functional configuration of the information processing device 1 according to the first embodiment.

[0024] As shown in FIG. 3, the information processing device 1 includes a creation procedure setting unit 25, a data receiving unit 30, a collaboration processing unit 31, a result processing unit 32, and a data management unit 33 as functional units realized on a processor (CPU 11).

[0025] The creation procedure setting unit 25 executes a setting process for setting the contents and procedures of the creation procedure manual. The setting process is set, for example, by a data user according to the purpose of using the output data. The data user communicates with the information processing device 1 through the data user terminal 3 and sets the creation procedure.

[0026] In this embodiment, the work procedure manual can be created without using any code such as a programming language by selecting and combining any block from a plurality of processing blocks prepared in advance. A data reception unit 30, a linkage processing unit 31, a result processing unit 32, a data management unit 33, etc. are registered as a plurality of processing blocks in the information processing device 1, and the data user can set the creation procedure manual by combining any processing block from the data reception unit 30, the linkage processing unit 31, the result processing unit 32, and the data management unit 33.

[0027] Next, the functions of the processing blocks that operate based on the created procedure manual will be described with reference to Fig. 4. Fig. 4 is a flow diagram showing the flow of the determination process by the information processing device 1 according to the first embodiment. Fig. 4 shows a work procedure manual including a data reception unit 30, a cooperation processing unit 31, and a result processing unit 32 that are set by a data user.

[0028] The data receiving unit 30 has a function of acquiring data to be processed from the target terminal 2 via a communication network. The data receiving unit 30 of this embodiment constructs a VA (Virtual Assistant) for receiving target data to be processed, such as text, voice, and image. The VA may be, for example, an AI chatbot using AI, or a device that accumulates messages exchanged in the past as knowledge and extracts and presents what is considered to be the most suitable answer to the target message from the accumulated knowledge, as an automatic response VA. Although further detailed explanation is omitted, the answer may include not only text, but also presentation of a URL specifying a web page, file, or distribution video, or attachment of a file, etc. The answer may be voice or an image (which may be a video) instead of text.

[0029] Fig. 5 is a diagram showing an example of target data input from the target terminal 2. Fig. 5 shows the contents of text input in a free format by a user in a chat message, email, or the like.

[0030] The collaboration processing unit 31 executes a connection process (connect function) for connecting to the generation AI server 101 of a vendor that is set in advance as a collaboration target, and an acquisition process (generation AI collaboration function) for outputting the target data acquired by the VA of the data receiving unit 30 and a script including commands to the generation AI server 101 and acquiring the output data output by the generation AI. Upon acquiring the target data, the collaboration processing unit 31 outputs the target data to the generation AI server 101 together with a preset prompt (script including commands), and acquires the output data generated by the generation AI server 101.

[0031] The generation AI server 101 to be linked in this example has a function of analyzing natural language using large-scale language models (LLMs) such as Chat-GPT. FIG. 6 is a diagram showing an example of output data acquired by the link processing unit 31. In this embodiment, a prompt is set in the link processing unit 31 so that the output data shown in FIG. 6 is output. In this example, the prompt is set so that "summary", "inquiry content", "positive / negative", "positive / negative score", "emotion strength", "emotion score", "subjectivity", "category", "category score", "FAQ", and "inquiry" are output from sentences included in the target data, which is text data.

[0032] The result processing unit 32 analyzes the output data based on a preset criterion, and executes a determination process to determine the next processing content to be performed on the output data. In the example of Fig. 6, the contents of "Positive / Negative" and "Category" are set as the criterion for determining the processing content, as they are considered to be the most important items among the items output in the output data.

[0033] The result processing unit 32 executes a judgment process to determine whether processing a or processing b is to be performed next on the output data based on the contents of "Positive / Negative" and "Category". When "Positive / Negative" is negative and "Category" is anger, the result processing unit 32 judges that the output data is of high business importance and determines processing a as the processing content. Processing a automatically sends an alert message to the data user terminal 3 of a manager such as a department manager, for example, and transitions to processing by the data management unit 33. Processing b transitions to processing by the data management unit 33 without automatically sending an alert message.

[0034] The result processing unit 32 is not limited to a configuration that determines whether or not an alert message should be sent for the output data. The result processing unit 32 may determine whether or not the output data should be stored in the database 90 based on a preset rule. In this way, the criteria for determining the output data can be appropriately determined depending on the purpose of use of the output data, etc.

[0035] Next, the data management unit 33 will be described with reference to Fig. 7. Fig. 7 is a flow diagram showing the flow of the storage process and presentation process of output data by the information processing device 1 according to the first embodiment.

[0036] As shown in FIG. 7, the data management unit 33 executes a storage process for storing the output data in the database 90 and a presentation process for presenting the stored output data to an administrator.

[0037] The storage process will now be described. After the process contents of the result processing unit 32 are determined, the data management unit 33 stores the output data in the database 90 in association with a user, such as an administrator, who uses the SaaS 102 and uses the output data.

[0038] The presentation process will now be described. When the data management unit 33 receives a request to view output data from a data user terminal 3 of an administrator who has the authority to use the output data, the data management unit 33 acquires the output data corresponding to the user from the database 90 and performs a process of presenting the output data to the data user terminal 3 that made the view request.

[0039] Next, an example in which conversations in a chat room for attendance reports are treated as target data to be processed will be described with reference to Fig. 8 and Fig. 9. Fig. 8 is a diagram showing an example of the conversation content in a chat room for attendance reports as target data. Fig. 9 is a diagram showing an example of output data after formatting created based on the conversation content in the chat room.

[0040] 8 is a function that enables chat among multiple people and is set for members who have the same SNS (Social Networking Service) application installed on the target terminal 2. The members belong to the same organization such as a company, belong to departments or projects that require mutual understanding of attendance records, and are required to report attendance schedules using the chat room as part of their work.

[0041] In the example of Figure 8, Yamada, one of the members, writes in the chat room, "I will be away from work tomorrow from 10:30 to 13:00 for business." Suzuki, another member, writes, "I will be taking the day off today due to poor health," and Sato, another member, writes, "6 / 27 11:00-13:00 away for medical treatment, 6 / 28 off for business, please understand." There is no uniformity in the way members report their attendance schedules, and even if you view the chat room through the target terminal 2, messages just flow by, making it difficult to immediately grasp the situation of each member.

[0042] In order to make it easier to understand the attendance schedule of members, the collaboration processing unit 31 sets a prompt to convert sentences including information such as the time of attendance, the time of absence, and whether or not the member is on vacation into simple words, and outputs the text content of the chat room received as target data via the data receiving unit 30 together with the prompt to the generation AI server 101. The collaboration processing unit 31 then obtains the output data converted into simple words by the generation AI server 101. The result processing unit 32 analyzes the output data input from the collaboration processing unit 31, formats it into a table format as shown in Fig. 9, and outputs it to the data management unit 33. In the example of Fig. 9, the attendance status indicating the time of attendance and absence of the member is shown in a table format for each date.

[0043] The data management unit 33 performs a process of storing the output data formatted in tabular form in the database 90 so that it can be confirmed on the member's target terminal 2, the data user terminal 3, etc. When the data management unit 33 receives a request to view the attendance schedule from the member's target terminal 2 or the data user terminal 3 of the administrator who manages the member's attendance, it executes a presentation process to present the attendance schedule based on the tabular output data to the target terminal 2 or data user terminal 3 that originated the viewing request. It is assumed that the authority to view the attendance schedule is set in advance for the member or administrator.

[0044] Next, an example of a combination of target data and output data different from those in Fig. 8 and Fig. 9 will be described with reference to Fig. 10. Fig. 10 is a diagram showing a schematic diagram of a state in which the contents of conversations between members of a chat room over a week as target data are registered in a database 90.

[0045] In the example shown in Fig. 10, the data receiving unit 30 acquires one week's conversation content in the chat room as target data, and the cooperation processing unit 31 cooperates with the generation AI server 101 to acquire a summary of the conversation content from the target data as output data. In this summarization process, for example, under the title of a one-week review, the key points of exchanges in each room, the main points of this week, candidates for action plans, etc. are extracted from the conversation content and output as output data. The output data is organized by the result processing unit 32 into a format suitable for data registration, and is registered in the database 90 by the data management unit 33.

[0046] 11 is a diagram showing a schematic diagram in which a summary of conversation content between members over the course of a week is registered in database 90, and each member is able to check the summary of conversation content over the week. As shown in FIG. 11, the summary of conversation content over the week obtained from the target data by the collaboration processing unit 31 in collaboration with the generation AI server 101 can be shared among members through the database 90.

[0047] The combination of the target data and the output data is not limited to the examples of FIGS. 5 and 6, 8 and 9, and 10 and 11. For example, the information processing device 1 can also analyze the requests of citizens. In this case, users with specific attributes, such as people living in a certain area, are asked to register in a chat room and tweet their requests (target data) in text, voice, or the like without being particularly conscious of the users. The data receiving unit 30 acquires the target data of the multiple target terminals 2 to be registered in the chat room. The collaboration processing unit 31 causes the generation AI server 101 to process summaries of each of the multiple target data to acquire output data. The result processing unit 32 converts the output data (summary of requests) into a format that is easy to analyze, such as a table format, depending on the purpose of the analysis. The data management unit 33 stores the output data after processing by the result processing unit 32 in the database 90 and visualizes the analysis results so that an administrator or the like who wants to check them can check them. In this way, the requests of citizens can be collected and analyzed using the target terminals 2 and communication networks used by each citizen without the hassle of telephone interviews or mailed questionnaires.

[0048] (1-1) As described above, the information processing device 1 of the first embodiment includes a data receiving unit 30 that executes a receiving process to receive target data to be processed, a collaboration processing unit 31 that executes an acquisition process to output a script including the target data and a command to a generation AI server 101 (a trained model including a large-scale language model) and acquire the output data output by the generation AI server 101, and a result processing unit 32 that executes a decision process to determine the processing content for the output data based on predetermined judgment criteria. This allows the user providing the target data to input the target data without needing to be aware of the existence of the generation AI server 101. Even if the user does not understand how the generation AI server 101 works, a prompt can be incorporated into the collaboration processing unit 31, allowing the data user to obtain output data in the format required. In this embodiment, the next process using the output data from the generation AI server 101 is determined by the result processing unit 32, so that processing after obtaining the output data can also be automatically branched. By making the processing of the result processing unit 32 rule-based, the process for utilizing the output data can be automated with simple processing.

[0049] In addition, the information processing device 1 of this embodiment executes a storage process to store the output data in the database 90, and further includes a data management unit 33 that is capable of executing a presentation process to present the output data stored by the storage process to an external terminal such as a data user terminal 3.

[0050] In a conventional system using a generation AI, even if an answer to a question can be obtained, the answer data must be stored separately. In this regard, according to the configuration of this embodiment, the output data of the generation AI server 101 is stored in the database 90 by a storage process and can be viewed from an external terminal such as the data user terminal 3, so that the searchability is good and the target data can be easily analyzed and visualized.

[0051] (1-5) In addition, the control method of the information processing device 1 of the first embodiment includes a reception step of executing a reception process to receive target data to be processed, a collaboration step of executing an acquisition process to output a script including the target data and a command to the generation AI server 101 and acquire the output data output by the generation AI server 101, and a decision step of executing a decision process to determine the processing content for the output data based on a predetermined judgment criterion. (1-6) In addition, the computer program of the first embodiment causes the computer to execute a reception step of executing a reception process to receive target data to be processed, a collaboration step of outputting a script including the target data and a command to the generation AI server 101 and executing an acquisition process to acquire the output data output by the generation AI server 101, and a decision step of executing a decision process to determine the processing content for the output data based on preset judgment criteria.

[0052] (2-1) In addition, the information processing device 1 of the first embodiment includes a data receiving unit 30 that executes a receiving process to receive target data to be processed, a collaboration processing unit 31 that outputs a script including the target data and a command to a generation AI server 101 (a trained model including a large-scale language model) and executes an acquisition process to acquire the output data output by the generation AI server 101, and a data management unit 33 that executes a storage process to store the output data in a database 90 and is capable of executing a presentation process to present the output data stored by the storage process to an external terminal such as a data user terminal 3. This allows the user providing the target data to input the target data without being aware of the existence of the generation AI server 101. Even if the user does not understand the mechanism of the generation AI server 101, a prompt can be incorporated into the collaboration processing unit 31, so that the data user can obtain output data in the format required. Furthermore, in a conventional system using a generation AI, even if an answer to a question can be obtained, the answer data must be stored separately. In this regard, in this embodiment, the output data of the generation AI server 101 is stored in the database 90 by a storage process and can be viewed from an external terminal such as the data user terminal 3, so that it is easy to search and statistical analysis and visualization of the target data can be easily performed.

[0053] Moreover, the information processing device 1 of this embodiment further includes a result processing unit 32 that executes a decision process for deciding the processing content for the output data based on a preset criterion. As a result, the next process using the output data of the generation AI server 101 is determined by the result processing unit 32, so that the process after obtaining the output data can also be automatically branched. By making the process of the result processing unit 32 rule-based, it is possible to automate the process for utilizing the output data with simple processing.

[0054] (2-4) In addition, the control method of the information processing device 1 of this embodiment includes a reception step of executing a reception process to receive target data to be processed, a collaboration step of executing an acquisition process to output a script including the target data and a command to the generation AI server 101 and acquire the output data output by the generation AI server 101, a storage step of executing a storage process to store the output data in the database 90, and a presentation step of executing a presentation process to present the output data stored by the storage process to an external terminal such as the data user terminal 3. (2-5) In addition, the computer program of the first embodiment includes a reception step of executing a reception process to receive target data to be processed, a collaboration step of executing an acquisition process to output a script including the target data and a command to the generation AI server 101 and acquire the output data output by the generation AI server 101, a storage step of executing a storage process to store the output data in the database 90, and a presentation step of executing a presentation process to present the output data stored by the storage process to an external terminal such as the data user terminal 3.

[0055] <Second embodiment> Next, an information processing device 1a according to a second embodiment, which has a different configuration from that of the first embodiment, will be described. Fig. 12 is a functional block diagram showing an example of the functional configuration of the information processing device 1a according to the second embodiment. Note that the same reference numerals are used for configurations common or similar to those of the first embodiment, and detailed descriptions thereof may be omitted.

[0056] As shown in FIG. 12, an information processing device 1a according to the second embodiment further includes an extraction processing unit 34 as a functional unit realized on a processor (CPU 11) in addition to the configuration according to the first embodiment.

[0057] The flow of processing by the information processing device 1a according to the second embodiment will be described with reference to Fig. 13. Fig. 13 is a flow diagram showing the flow of extraction processing and link processing by the information processing device 1a according to the second embodiment.

[0058] When the data receiving unit 30 receives a question as target data, it outputs the target data to the extraction processing unit 34. The extraction processing unit 34 accesses the database 90 to obtain information related to the target data received by the data receiving unit 30.

[0059] The database 90 in the second embodiment is a knowledge database in which information on target data to be processed is registered. The information on the target data includes internal information that is not disclosed on the Internet, such as business know-how and past implementation measures. For example, improvement actions A to E set for various issues are registered in the database 90, and among these, improvement actions A and E are assumed to have been implemented or planned in the past to improve column efficiency. When the question content as the target data is related to column efficiency, the extraction processing unit 34 outputs improvement actions A and E as extracted data together with the target data to the collaboration processing unit 31.

[0060] The classification of the target data by the extraction processing unit 34 may be performed based on rules using a thesaurus, or a learning model such as machine learning may be used. The improvement measures may be an action plan included in the conversation content for one week registered when the first embodiment is described with reference to Figs. 10 and 11. The database 90 to be extracted by the extraction processing unit 34 may not be one, but may be multiple. In this case, the extraction processing unit 34 may perform a process of selecting the database 90 to be extracted from the multiple databases according to the contents of the target data.

[0061] The collaboration processing unit 31 outputs the question content and the extracted data to the generation AI server 101 along with a preset prompt, and obtains output data. This output data reflects past accumulated information related to the question content in the database 90. In this example, improvement actions A and E are output to the generation AI server 101 along with the question content, and output data incorporating improvement actions A and E is generated.

[0062] The result processing unit 32 outputs the answer as output data to the subject who asked the question via the VA of the data receiving unit 30. Although omitted in Fig. 13, the data management unit 33 may perform a process of storing the answer as output data in the database 90 together with the question content as the subject data.

[0063] (2-2) As described above, the information processing device 1b of the second embodiment further includes an extraction processing unit 34 that executes an extraction process to extract extracted data related to the target data from the database 90 in which information related to the target data is registered, and the extraction processing unit 34 outputs the extracted data to the collaboration processing unit 31. This allows information such as know-how that is not publicly available on the Internet to be stored in database 90, making it possible to reflect that information in the output processing of the generation AI. Since text data can be registered in database 90, a system that reflects stored information in the generation AI can be constructed without preparing an advanced environment for building a vector database that digitizes text data, such as RAG (Retrieval Augmented Generation).

[0064] (2-3) In addition, in this embodiment, the cooperative processing unit 31 outputs the target data and a script including a command together with the extracted data to the generation AI server 101, and obtains the output data. This makes it possible to reflect extracted data in the output data output by the generation AI through the simple process of combining target data consisting of inquiries and other questions with extracted data such as solution measures related to the target data, without complicating the processing.

[0065] <Third embodiment> Next, an example of creating output data using a plurality of types of generation AI servers 101a to 101c will be described with reference to Fig. 14. Fig. 14 is a flow diagram showing the flow of a plurality of cooperative processes by an information processing device 1b according to a third embodiment.

[0066] 14, an information processing device 1b according to the third embodiment includes a first link processing unit 31a, a second link processing unit 31b, and a third link processing unit 31c. Target data according to the third embodiment is audio data such as recordings of meetings and interviews.

[0067] The flow of processing in the third embodiment will be described below. First, the data receiving unit 30 acquires voice data via the VA, and outputs the data to the first cooperation processing unit 31a.

[0068] The first collaboration processor 31a is preset with a generation AI server 101a of vendor X, which specializes in a function for converting voice to text, as a collaboration target. The first collaboration processor 31a outputs voice data to the generation AI server 101a of vendor X, and acquires text data as output data resulting from the character conversion. The first collaboration processor 31a outputs the acquired text data to the second collaboration processor 31b.

[0069] The second collaboration processing unit 31b is preset with a generation AI server 101b of vendor Y, which specializes in the function of summarizing text data, as a collaboration target. The generation AI server 101b is an LLM. The second collaboration processing unit 31b outputs text data to the generation AI server 101b of vendor Y, and obtains output data that has been subjected to the summary processing. The second collaboration processing unit 31b outputs the obtained summary data to the third collaboration processing unit 31c.

[0070] The third collaboration processing unit 31c is preset with a vendor Z generation AI server 101c, which is specialized in the function of performing industry dictionary analysis on text, as a collaboration target. The third collaboration processing unit 31c outputs summary data to the vendor Z generation AI server 101c, and acquires output data after industry dictionary analysis. The third collaboration processing unit 31c outputs the acquired output data after industry dictionary analysis to the result processing unit 32.

[0071] The result processing unit 32 processes the output data after the industry dictionary analysis based on preset processing contents, and outputs the data to the data management unit 33. For example, the result processing unit 32 converts the output data after the industry dictionary analysis into a format suitable for storage, and the data management unit 33 stores the output data after processing by the result processing unit 32 in the database 90. As a result, the output data stored in the database 90 is ready to be presented in response to a request from a data user.

[0072] (1-2) As described above, in the information processing device 1b of the third embodiment, multiple collaboration processing units (first collaboration processing unit 31a, second collaboration processing unit 31b, and third collaboration processing unit 31c) each execute an acquisition process in collaboration with a different generation AI server 101a to 101c. This allows the target data to be converted into output data with high accuracy by setting the generation AI servers 101a-101c according to the characteristics of each different process. By using multiple types of generation AI servers 101a-101c, output data in various formats can be output, and various requests from data users can be met.

[0073] <Fourth embodiment> Next, an information processing device 1c according to a fourth embodiment having a creation procedure setting unit 25 configured differently from the creation procedure setting unit 25 according to the third embodiment will be described with reference to Fig. 15. Fig. 15 is a functional block diagram showing an example of the functional configuration of the information processing device 1c according to the fourth embodiment.

[0074] As shown in FIG. 15, the information processing device 1c of the fourth embodiment has a creation procedure setting unit 25a, a data receiving unit 30, a collaboration processing unit 31, a result processing unit 32, and a data management unit 33 as functional units realized on a processor (CPU 11).

[0075] In the fourth embodiment, the creation procedure setting unit 25a has an analysis function for analyzing the processing required for the work procedure manual based on the request information input by the data user, and a proposal function for proposing the generation AI servers 101a to 101c required for the required processing. The data receiving unit 30, the collaboration processing unit 31, the result processing unit 32, and the data management unit 33 in the fourth embodiment have the same configurations as those described in the above embodiments. The creation procedure setting unit 25a is applicable to any of the information processing devices 1, 1a to 1b in the first to third embodiments.

[0076] An example of a process for creating a work procedure manual based on request information input by a data user will be described with reference to Fig. 16. Fig. 16 is a flowchart showing an example of the flow of a process for creating a work procedure manual by the information processing device 1c according to the fourth embodiment.

[0077] In step S1, the data receiving unit 30 acquires request information including the content of output data ultimately desired to be obtained from the data user, the purpose of data processing, etc. The data receiving unit 30 receives the request information via the data user terminal 3, etc., for example, in the form of a chat using VA, text such as e-mail, voice input, etc.

[0078] In step S2, the creation procedure setting unit 25a executes an analysis process that analyzes the request information acquired by the data receiving unit 30 and creates a plurality of processes for obtaining output data requested by the data user from the request information. The method of the analysis process is not particularly limited. For example, the analysis process may be in a form of extracting a work procedure manual that matches the data user from among a plurality of processes that are registered in advance by a rule base using keywords or the like, or may be in a form of selecting a work procedure manual that includes a plurality of processes that corresponds to the one selected by the data user from among options prepared in advance. Alternatively, the analysis process may be in a form that uses a generation AI such as machine learning or deep learning.

[0079] In step S3, the creation procedure setting unit 25a sets the generation AI server 101 to be linked depending on the multiple process types created by the analysis. The creation procedure setting unit 25a, for example, refers to a database 90 in which multiple generation AI servers 101 corresponding to the process types and their connection information are registered in advance, and extracts a generation AI server 101 appropriate for the process type.

[0080] In step S4, the creation procedure setting unit 25a executes a presentation process to present to the data user a creation procedure manual including multiple processes for obtaining output data and a generation AI server 101 corresponding to each process. The data user sets the creation procedure manual based on the presented multiple processes and the generation AI server 101 corresponding to each process. The data user may use the presented creation procedure manual as is, or may modify the created creation procedure manual and use it.

[0081] (1-3) As described above, the information processing device 1c of the fourth embodiment further includes a creation procedure setting unit 25a that sets the content and procedure of processing for target data. The creation procedure setting unit 25a analyzes request information indicating requests input by a data user, and executes an analysis process of generating, in a plurality of blocks, a process for obtaining the output data that satisfies the request information. This automatically generates processing divided into multiple blocks based on the requested information. Data users can use the generated multiple blocks as is or refer to them, effectively reducing the effort and time required to create a procedure manual. It also prevents situations where data users are not familiar with the procedure manual and try to force complex processing into a single process, resulting in the desired output data not being generated.

[0082] (1-4) In addition, in this embodiment, the creation procedure setting unit 25a executes a proposal process that proposes candidates for the generation AI servers 101a to 101c corresponding to each of a plurality of processes. As a result, even if it is necessary to cooperate with the generation AI server 101 for each of multiple types of processing, the generation procedure setting unit 25a presents the generation AI servers 101a-c suitable for the processing, thereby reducing the effort and time required to create a generation procedure manual. Even if the data user who sets the generation procedure manual is not familiar with the areas of expertise of the generation AI servers 101a-c, it is possible to create a generation procedure manual according to the characteristics of each generation AI server 101a-c.

[0083] 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 that can achieve the object of the present invention are included in the present invention.

[0084] Moreover, the above-mentioned series of processes can be executed by hardware or software. In other words, the above-mentioned functional configuration is merely an example and is not particularly limited. That is, it is sufficient that the information processing device 1, 1a to 1c has a function capable of executing the above-mentioned series of processes as a whole, and the type of functional block used to realize this function is not particularly limited to the above-mentioned example. Furthermore, the location of the functional block is not particularly limited and may be arbitrary. For example, the functional block of the information processing device 1, 1a to 1c may be transferred to another device or the like. Conversely, the functional block of the other device may be transferred to a server or the like. Furthermore, one functional block may be configured by a single piece of hardware, a single piece of software, or a combination thereof.

[0085] When a series of processes is executed by software, a program constituting the software is installed in a computer or the like from a network or a recording medium. The computer may be a computer built 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.

[0086] The recording medium containing such a program may be configured not only as a removable medium (not shown) that is distributed separately from the device main body in order to provide the program, but also as a recording medium provided in a state in which it is pre-installed in the device main body, etc. Since the program can be distributed via a network, the recording medium may be mounted on or accessible to a computer that is connected or connectable to the network.

[0087] In this specification, the steps describing the program recorded on the recording medium include not only processes performed in chronological order, but also processes that are not necessarily performed in chronological order but are performed in parallel or individually. In addition, in this specification, the term "system" refers to an overall device composed of multiple devices, multiple means, etc. [Explanation of symbols]

[0088] 1, 1a to 1c Information processing device 2. Target device 3 Data user terminal 25, 25a Work procedure setting section 30 Data Reception Department 31 Collaboration Processing Unit 31a First cooperation processing unit 31b Second cooperation processing unit 31c Third link processing unit 32 Result Processing Section 33 Data Management Department 34 Extraction processing section 100 Data Processing System 101, 101a~101c Generation AI server

Claims

1. a data receiving unit that executes a receiving process for receiving target data to be processed; A cooperation processing unit that executes an acquisition process of outputting a script including the target data and a command to a trained model including a large-scale language model and acquiring output data output by the trained model; a data management unit that executes a storage process for storing the output data in a database and executes a presentation process for presenting the output data stored by the storage process on an external terminal; An information processing device comprising:

2. An extraction processing unit that executes an extraction process to extract extraction data related to the target data from a database in which information related to the target data is registered, The extraction processing unit outputs the extracted data to the cooperation processing unit. The information processing device according to claim 1 .

3. The collaboration processing unit outputs a script including the target data and a command together with the extracted data to the trained model, and acquires the output data. The information processing device according to claim 2 .

4. a receiving step of executing a receiving process for receiving target data to be processed; A linking step of executing an acquisition process of outputting a script including the target data and a command to a trained model including a large-scale language model and acquiring output data output by the trained model; a storing step of executing a storing process for storing the output data in a database; a presentation step of executing a presentation process for presenting the output data stored by the storage process on an external terminal; A method for controlling an information processing device comprising:

5. a receiving step of executing a receiving process for receiving target data to be processed; A linking step of executing an acquisition process of outputting a script including the target data and a command to a trained model including a large-scale language model and acquiring output data output by the trained model; a storing step of executing a storing process for storing the output data in a database; a presentation step of executing a presentation process for presenting the output data stored by the storage process on an external terminal; A program for causing a computer to execute the following.

Images