Information processing system

Abstract

The present invention aims to suppress a usage fee of a large-scale language model. An information processing system includes a storage unit that stores first chat data related to a first chat of a user and a large-scale language model, an acquisition unit that acquires a question sentence input by the user in a second chat of the user and the large-scale language model started after the first chat, and an output unit that outputs an answer to the acquired question sentence as an answer to the corresponding question sentence included in the first chat data when the corresponding question sentence is included in a plurality of utterances included in the first chat data.

Description

Technical Field

[0001] This invention relates to the technical field of an information processing system. Background Technology

[0002] As such a system, for example, a system has been proposed that enables a language model to generate query data based on a document, and uses the document-query data pair for learning a retrieval model for a chatbot (see Patent Document 1).

[0003] Patent Document 1: Japanese Patent Application Publication No. 2023-076413 Summary of the Invention

[0004] As a chatbot, a chatbot utilizing Large Language Models (LLMs) has been proposed. The length of a series of chats that a chatbot can record is limited. Therefore, when the length of a series of chats reaches the limit, the user must restart the chat. As a result, for example, the user may re-enter the same question as previously asked. Utilizing large-scale language models typically incurs a usage fee. Therefore, users sometimes incur additional usage fees by re-entering the same question as previously asked. Furthermore, large-scale language models refer to language models built using very large datasets and deep learning techniques.

[0005] The present invention was made in view of the above-mentioned problems, and its objective is to provide an information processing system capable of suppressing the usage costs of large-scale language models.

[0006] An information processing system according to one aspect of the present invention comprises: a storage unit that stores first chat data related to a first chat between a user and a large-scale language model; an acquisition unit that acquires a question input by the user in a second chat between the user and the large-scale language model after the first chat; and an output unit that, when the first chat data includes a question corresponding to the acquired question among a plurality of utterances, outputs the answer to the corresponding question included in the first chat data as an answer to the acquired question. Attached Figure Description

[0007] Figure 1 This is a diagram showing the structure of the information processing system involved in the implementation method.

[0008] Figure 2 This is a block diagram illustrating an example of the structure of the computing device involved in the implementation.

[0009] Figure 3 This is a flowchart illustrating the operation of the information processing apparatus involved in the implementation method. Detailed Implementation

[0010] refer to Figures 1 to 3 The implementation methods involved in the information processing system are described. Figure 1 In this system, information processing system 1 includes information processing device 10, server 20, and knowledge base 30. Information processing device 10, server 20, and knowledge base 30 are configured to communicate with each other via a network NW. Server 20 is a server used for applying Large Scale Language Model (LLM). Therefore, server 20 can be called an LLM server. Alternatively, server 20 can be a cloud server.

[0011] Multiple text data items can be registered in knowledge base 30. This text data can be fragmented data, representing the text contained in a document. Such fragmented data can be called "chunks." Furthermore, methods for segmenting the text contained in a document can include, for example, segmenting with a constant length (in other words, fixed length), segmenting by sentence based on sentence delimiters, or segmenting based on structures such as Markdown. Additionally, knowledge base 30 can vectorize and register multiple text data items separately. That is, knowledge base 30 can be a vector database / vector storage. Besides text data, image data can also be registered in knowledge base 30.

[0012] exist Figure 1 The information processing device 10 includes a computing unit 11, a storage unit 12, a communication unit 13, an input unit 14, and an output unit 15. The computing unit 11, storage unit 12, communication unit 13, input unit 14, and output unit 15 are connected via a data bus 16. Furthermore, the information processing device 10 can be a personal computer, a tablet terminal, or a smartphone.

[0013] The arithmetic unit 11 may have a processor. Furthermore, the arithmetic unit 11 may have a single processor or multiple processors. That is, the arithmetic unit 11 may have more than one processor. Additionally, the processor may be a multi-core processor. In the case where the arithmetic unit 11 has a single processor that functions as a multi-core processor, it can be said that the arithmetic unit 11 logically has multiple processors.

[0014] The processor may be at least one of a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), a Field Programmable Gate Array (FPGA), and a Tensor Processing Unit (TPU).

[0015] Storage device 12 may be at least one of random access memory (RAM), read-only memory (ROM), hard disk drive, magneto-optical disk drive, solid-state drive (SSD), and optical disk array. That is, storage device 12 may be implemented by a single device or by multiple devices.

[0016] The communication device 13 can communicate with external devices of the information processing device 10. In addition, the communication device 13 can perform wired communication or wireless communication.

[0017] Input device 14 is a device capable of accepting information input from an external source to information processing device 10. Input device 14 may include user-operable devices (e.g., keyboard, mouse, touch panel, etc.) of information processing device 10. Input device 14 may include, for example, a recording medium reading device capable of reading information recorded on a recording medium removable from information processing device 10, such as a Universal Serial Bus (USB) memory. Furthermore, when information is input to information processing device 10 via communication device 13 (in other words, when information processing device 10 obtains information via communication device 13), communication device 13 may function as an input device.

[0018] Output device 15 is a device capable of outputting information to the outside of information processing device 10. Output device 15 includes a display device 151 capable of outputting visual information such as characters and images. Additionally, output device 15 may include a speaker capable of outputting auditory information such as sound. Output device 15 may include a vibration motor capable of outputting tactile information such as vibration. Output device 15 may include a printer. Output device 15 can output information to recording media removable from information processing device 10, such as a USB memory. Furthermore, when information processing device 10 outputs information via communication device 13, communication device 13 can function as an output device.

[0019] Storage device 12 is capable of storing desired data. The computer program CP executed by the arithmetic unit 11 can be stored in storage device 12. When the arithmetic unit 11 executes the computer program CP, storage device 12 can temporarily store data temporarily used by the arithmetic unit 11.

[0020] Furthermore, the computer program CP can be recorded on a computer-readable and non-temporary recording medium. In this case, the computer program CP can be stored in the storage device 12 by reading the recording medium using a recording medium reading device (not shown) included in the information processing device 10. Additionally, at least one of optical discs, magnetic media, magneto-optical discs, semiconductor memory, and any medium capable of storing other programs can be used as the recording medium. Furthermore, the computer program CP can also be obtained from an external (not shown) device outside the information processing device 10 via the communication device 13. In other words, the computer program CP can be downloaded from an external device to the storage device 12 of the information processing device 10.

[0021] The arithmetic unit 11 (e.g., a processor) can perform the processing to be performed by the information processing unit 10 together with the storage device 12 storing the computer program CP (in other words, together with the storage device 12 and the computer program CP stored in the storage device 12). For example, the arithmetic unit 11 can implement logical function blocks for performing the processing to be performed by the information processing unit 10 within the arithmetic unit 11 (e.g., within the processor) by executing the computer program CP.

[0022] (Chatbot)

[0023] Server 20 and knowledge base 30 can provide chatbot services using Retrieval-Augmented Generation (RAG). The computer program CP stored in storage device 12 includes a computer program (hereinafter, appropriately referred to as the "chat application") related to the application used to utilize the aforementioned chatbot service. Users of information processing device 10 can utilize the chatbot service by launching the chat application via input device 14.

[0024] For example, a user can enter a question in the input field of a chat application via input device 14. Additionally, users are not limited to character input; they can also use voice input. A "question" is not limited to interrogative sentences. For example, a "question" can be a sentence containing requests, instructions, or commands, such as "Please tell me about ****" or "Please answer about ****". Therefore, a "question" is not limited to interrogative sentences; it includes sentences containing requests, instructions, or commands. That is, a "question" can refer to a sentence requesting an answer from the other party.

[0025] The processing unit 11 can retrieve information from the knowledge base 30 based on the input question. The processing unit 11 can send first information representing the input question and the search results to the server 20 via the communication device 13. Upon receiving the first information, the server 20 can input a prompt including the question and search results represented by the first information into a large-scale language model. The server 20 can obtain an answer to the question generated by the large-scale language model. The server 20 can send second information representing the answer to the information processing unit 10. Upon receiving the second information, the processing unit 11 of the information processing unit 10 can control the display device 151 to display the answer represented by the second information.

[0026] However, in chatbot services, the length of a series of chats that can be recorded is limited. Therefore, if the length of a series of chats reaches the prescribed limit, the user must restart the chat. Here, the chat application according to this embodiment has the function of storing chat data (in other words, chat history) related to the user's chat with a large-scale language model in the chatbot service as part of its functionality.

[0027] In the information processing apparatus 10, when the length of a series of chats reaches a predetermined limit, for example, a chat application (specifically, the computing device 11) can associate the chat data related to the series of chats with the user and store it in the storage device 12. Alternatively, the chat data can be stored in a different device than the storage device 12 (e.g., the server 20). Furthermore, the computing device 11 processes the chat data related to the current chat and the chat data related to new chats as different data.

[0028] If a series of chats reaches the specified length limit and the user restarts the chat, previous chat content is typically not displayed. Therefore, users may sometimes re-enter the same question as before. In this case, additional usage fees may be incurred for the chatbot service.

[0029] In this embodiment, in order to suppress the usage fee of chatbot services, the computing device 11 includes an acquisition unit 111, a retrieval unit 112, an output unit 113, and an input unit 114 (see reference). Figure 2 The acquisition unit 111, retrieval unit 112, output unit 113, and input unit 114 can be implemented as the aforementioned logic function blocks. Alternatively, at least one of the acquisition unit 111, retrieval unit 112, output unit 113, and input unit 114 can be implemented as a physical processing circuit. At least one of the acquisition unit 111, retrieval unit 112, output unit 113, and input unit 114 can be implemented in a manner that combines logic function blocks and physical processing circuits.

[0030] refer to Figure 3 The flowchart below explains the operation of the information processing device 10. As a prerequisite, the storage device 12 of the information processing device 10 stores chat data related to past chats. The user can enter a question in the input field of the chat application via the input device 14. At this time, the acquisition unit 111 of the processing unit 11 acquires the entered question (step S101).

[0031] The retrieval unit 112 of the processing device 11 retrieves chat data (in other words, chat history) stored in the storage device 12 based on the question obtained in step S101 (step S102). In step S102, the retrieval unit 112 retrieves a question corresponding to the obtained question. Furthermore, "corresponding question" can refer to a question that is the same as or similar to the obtained question. In step S102, the retrieval unit 112 can calculate the cosine similarity between the obtained question and each of the multiple utterances included in the chat data stored in the storage device 120. The retrieval unit 112 can retrieve a question corresponding to the obtained question from the chat data based on the cosine similarity. However, cosine similarity is an example and not limited to this.

[0032] The retrieval unit 112 determines whether the multiple utterances included in the chat data include a question sentence corresponding to the aforementioned acquired question sentence (step S103). In the processing of step S103, if it is determined that the corresponding question sentence is included (step S103: Yes), the output unit 113 of the processing device 11 retrieves the answer to the corresponding question sentence from the chat data stored in the storage device 12 (step S104). The output unit 113 outputs the retrieved answer as an answer to the acquired question sentence (i.e., the question sentence entered by the user this time) (step S105). At this time, the output unit 113 can control the display device 151 to display the retrieved answer. Furthermore, in this case, the processing device 11 may not perform a search of the knowledge base 30.

[0033] In step S103, if it is determined that the corresponding question sentence is not included (step S103: No), the input unit 114 of the computing device 11 sends the question sentence obtained in step S101 to the server 20 (step S106). As a result, the question sentence input by the user is input into the large-scale language model. Furthermore, the computing device 11 can retrieve information from the knowledge base 30 based on the question sentence input by the user. In step S106, the input unit 114 can send information representing the question sentence input by the user and the retrieval results from the knowledge base 30 (equivalent to the first information mentioned above) to the server 20.

[0034] (Technical effect)

[0035] In information processing system 1, when the chat history includes a question corresponding to the question currently input by the user, an answer to the corresponding question is retrieved from the chat history. Then, this retrieved answer is output as an answer to the currently input question. That is, in information processing system 1, when the chat history includes a question corresponding to the question currently input by the user, a large-scale language model is not used. Therefore, information processing system 1 according to this embodiment can suppress the cost of using a large-scale language model.

[0036] The invention described below is an embodiment derived from the above-described implementation.

[0037] One aspect of the invention relates to an information processing system comprising: a storage unit storing first chat data related to a first chat between a user and a large-scale language model; an acquisition unit acquiring a question input by the user in a second chat between the user and the large-scale language model after the first chat; and an output unit outputting, when the first chat data includes a question corresponding to the acquired question among a plurality of utterances, an answer to the corresponding question included in the first chat data as an answer to the acquired question. In the above embodiment, "storage device 12" is equivalent to an example of "storage unit," "acquisition unit 111" is equivalent to an example of "acquisition unit," and "output unit 113" is equivalent to an example of "output unit."

[0038] The information processing system described above may include an input unit. If the input unit does not include a question sentence corresponding to the acquired question sentence among the multiple utterances included in the first chat data, it inputs the acquired question sentence into the large-scale language model. In the above embodiment, "input unit 114" is an example of an "input unit".

[0039] This invention is not limited to the embodiments described above, and appropriate modifications can be made without departing from the spirit or concept of the invention as read in its entirety from the claims and description. Information processing systems that accompany such modifications are also included within the technical scope of this invention.

[0040] Symbol Explanation

[0041] 10-Information processing device, 11-Arithmetic processing device, 12-Storage device, 13-Communication device, 14-Input device, 15-Output device, 111-Acquisition unit, 112-Retrieval unit, 113-Output unit, 114-Input unit, 151-Display device.

Claims

1. An information processing system, characterized in that, have: A storage unit that stores the first chat data related to the first chat between the user and the large-scale language model; The acquisition unit, which, after the first chat, acquires the question input by the user during the second chat between the user and the large-scale language model; and The output unit, when the first chat data includes a question corresponding to the acquired question among the multiple utterances, outputs the answer to the corresponding question included in the first chat data as an answer to the acquired question.

2. The information processing system according to claim 1, characterized in that, have: An input unit, which, if the multiple utterances included in the first chat data do not include a question sentence corresponding to the acquired question sentence, inputs the acquired question sentence into the large-scale language model.

Images