Information processing device

Abstract

To improve the response accuracy of large-scale language models. [Solution] The information processing device (10) includes a search means (111) that searches a database (30) based on a question entered by a user, an input means (112) that inputs text data as part of the search results from the database by the search means and the question into a large-scale language model, an acquisition means (113) that obtains the answer to the question generated by the large-scale language model, a display means (151) that displays the answer, and a reception means (14) that can accept user input. When the reception means receives user input indicating that it selects a displayed answer, the display means displays the search results.

Description

Technical Field

[0001] The present invention relates to the technical field of information processing apparatuses.

Background Art

[0002] As an apparatus of this type, for example, an apparatus has been proposed that generates query data based on a document for a language model and uses a pair of the document and the query data for learning a search model for a chatbot (see Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] As a chatbot, by combining a large language model (Large Language Models: LLM) and the search of a specific information source (hereinafter, appropriately referred to as a "knowledge base"), a chatbot using a mechanism (Retrieval - Augmented Genration: RAG) that gives an independent information source to the large language model has been proposed. In a chatbot using RAG, by inputting a part of the search results of the knowledge base into the large language model, an independent information source may be given to the large language model. At this time, the information expected by the user may not be input into the large language model. As a result, the answer generated by the large language model may be different from the answer sought by the user. Note that a large language model is a language model constructed using a very large dataset and deep learning technology.

[0005] This invention has been made in view of the above circumstances, and aims to provide an information processing device that can improve the response accuracy of large-scale language models. [Means for solving the problem]

[0006] An information processing device according to one aspect of the present invention comprises: a search means for searching a database based on a question sentence entered by a user; an input means for inputting text data as part of the search results from the database obtained by the search means and the question sentence into a large-scale language model; an acquisition means for obtaining an answer to the question sentence generated by the large-scale language model; a display means for displaying the answer; and a receiving means capable of receiving user input, wherein when the receiving means receives user input indicating that the user selects the displayed answer, the display means displays the search results. [Brief explanation of the drawing]

[0007] [Figure 1] This is a diagram showing the configuration of an information processing system according to an embodiment. [Figure 2] This block diagram shows an example of the configuration of a computing device according to the embodiment. [Figure 3] This figure shows an example of a displayed image. [Figure 4] This figure shows other examples of the displayed image. [Modes for carrying out the invention]

[0008] Embodiments relating to the information processing device will be described with reference to Figures 1 to 4. In Figure 1, the information processing system 1 comprises an information processing device 10, a server 20, and a knowledge base 30. The 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 for operating a large-scale language model (LLM). For this reason, server 20 may be referred to as an LLM server. Server 20 may be a cloud server.

[0009] Knowledge Base 30 may contain multiple text data entries. Each text data entry may be fragmented data generated by splitting a single document. Such fragmented data may be referred to as "chunks." Specific examples of methods for splitting a single document include splitting at fixed lengths, splitting at sentence units based on sentence delimiters, and splitting based on structure such as Markdown. Each text data entry may be vectorized text data. In other words, Knowledge Base 30 may be a vector database / vector store.

[0010] (Information processing device 10) In Figure 1, the information processing device 10 comprises an arithmetic unit 11, a storage device 12, a communication device 13, an input device 14, and an output device 15. The arithmetic unit 11, storage device 12, communication device 13, input device 14, and output device 15 are connected via a data bus 16. The information processing device 10 may be a personal computer, a tablet terminal, or a smartphone.

[0011] The arithmetic unit 11 may have a processor. The arithmetic unit 11 may have a single processor or multiple processors. In other words, the arithmetic unit 11 may have one or more processors. Furthermore, the processor may be a multi-core processor. If the arithmetic unit 11 has a single processor that is a multi-core processor, then logically, the arithmetic unit 11 can be said to have multiple processors.

[0012] The processor may be at least one of the following: CPU (Central Processing Unit), GPU (Graphics Processing Unit), FPGA (Field Programmable Gate Array), and TPU (Tensor Processing Unit).

[0013] The storage device 12 may be at least one of the following: RAM (Random Access Memory), ROM (Read Only Memory), hard disk drive, magneto-optical disk drive, SSD (Solid State Drive), and optical disk array. In other words, the storage device 12 may be implemented by a single device or by multiple devices.

[0014] The communication device 13 may be capable of communicating with an external device (for example, a server 20) of the information processing device 10. The communication device 13 may use either wired or wireless communication.

[0015] The input device 14 is a device capable of receiving information input to the information processing device 10 from an external source. The input device 14 may include an operating device (e.g., a keyboard, mouse, touch panel, etc.) that can be operated by the user of the information processing device 10. The input device 14 may include a recording medium reader capable of reading information recorded on a recording medium that can be attached to and detached from the information processing device 10, such as a USB (Universal Serial Bus) memory. When information is input to the information processing device 10 via the communication device 13 (in other words, when the information processing device 10 acquires information via the communication device 13), the communication device 13 may function as an input device.

[0016] The output device 15 is a device capable of outputting information to the outside of the information processing device 10. The output device 15 has a display device 151 capable of outputting visual information such as characters and images as the above information. The output device 15 may also have a speaker capable of outputting auditory information such as sound as the above information. The output device 15 may also have a vibration motor capable of outputting tactile information such as vibration as the above information. The output device 15 may also have a printer. The output device 15 may be capable of outputting information to a recording medium that can be attached to and detached from the information processing device 10, such as a USB memory stick. When the information processing device 10 outputs information via the communication device 13, the communication device 13 may function as an output device.

[0017] The storage device 12 is capable of storing desired data. The storage device 12 may store the computer program CP that the arithmetic unit 11 will execute. The storage device 12 may temporarily store data that the arithmetic unit 11 will use temporarily when the arithmetic unit 11 is executing the computer program CP.

[0018] Furthermore, the computer program CP may be recorded on a non-temporary recording medium that is readable by a computer. In this case, the computer program CP may be stored in the storage device 12 by reading the recording medium using a recording medium reading device (not shown) provided by the information processing device 10. Furthermore, at least one of the following may be used as the recording medium: an optical disc, a magnetic medium, a magneto-optical disc, a semiconductor memory, and any other medium capable of storing a program. Furthermore, the computer program CP may be obtained from an external device (not shown) of the information processing device 10 via a communication device 13. In other words, the computer program CP may be downloaded from an external device to the storage device 12 of the information processing device 10.

[0019] The arithmetic unit 11 (for example, a processor) may execute the processing that the information processing device 10 should perform together with the memory device 12 in which the computer program CP is stored (in other words, together with the memory device 12 and the computer program CP stored in the memory device 12). For example, by the arithmetic unit 11 executing the computer program CP, a logical functional block for executing the processing that the information processing device 10 should perform may be realized within the arithmetic unit 11 (for example, within the processor).

[0020] The server 20 and the knowledge base 30 provide a chatbot service using RAG. The arithmetic unit 11 of the information processing device 10 has a search unit 111, an input unit 112, an acquisition unit 113, and a modification unit 114 for using the chatbot service (see FIG. 2). The search unit 111, the input unit 112, the acquisition unit 113, and the modification unit 114 may be realized as the above-described logical functional blocks. Note that at least one of the search unit 111, the input unit 112, the acquisition unit 113, and the modification unit 114 may be realized as a physical processing circuit. At least one of the search unit 111, the input unit 112, the acquisition unit 113, and the modification unit 114 may be realized in a form in which logical functional blocks and physical processing circuits are mixed.

[0021] For example, the user may use the chatbot service via the information processing device 10. In this case, the user may input a question sentence via the input device 14 of the information processing device 10. Here, the "question sentence" is not limited to an interrogative sentence. For example, the "question sentence" may be a sentence including expressions such as "Tell me about ****", "Answer ****", etc., including requests, instructions, commands, etc. Therefore, the "question sentence" is not limited to an interrogative sentence, and is a concept including, for example, a sentence including an expression such as a request, an instruction, or a command. That is, the "question sentence" may mean a sentence that requests an answer from the other party.

[0022] The search unit 111 of the information processing apparatus 10 may search the knowledge base 30 based on the question sentence. For example, the search unit 111 may calculate a search score indicating the degree of association between the question sentence and the text data registered in the knowledge base 30. The search unit 111 may extract text data having a search score equal to or higher than a predetermined value as text data related to the question sentence. The search unit 111 may output a search result including the extracted text data and its search score. That is, in the search result, the text data and its search score may be associated with each other. Note that the search unit 111 may extract a plurality of text data related to the question sentence. That is, the search unit 111 may extract one or more text data related to the question sentence from the knowledge base 30. Note that various existing modes can be applied to the calculation method of the search score. Therefore, a detailed description of the calculation method of the search score is omitted.

[0023] The input unit 112 of the information processing apparatus 10 transmits a prompt including the question sentence and text data as part of the search result to the server 20 via the communication device 13. As a result, the prompt is input into the large language model. Here, the text data included in the prompt may be a predetermined number of text data (for example, the top 5 text data in terms of search score) having a relatively high search score among the text data included in the search result. That is, not all of the text data included in the search result may be included in the prompt.

[0024] The server 20 transmits an answer to the question sentence generated by the large language model to the information processing apparatus 10. The acquisition unit 113 of the information processing apparatus 10 acquires the answer when the information processing apparatus 10 receives the answer generated by the large language model. The arithmetic unit 11 of the information processing apparatus 10 controls the display device 151 to display the answer. As a result, the image 210 shown in FIG. 3 may be displayed on the display device 151. For example, the answer generated by the large language model may be displayed in the area 211 of the image 210.

[0025] Incidentally, the answer generated by the large-scale language model may differ from the answer the user is seeking. In such cases, in this embodiment, if the user selects an answer generated by the large-scale language model included in image 210 via the input device 14 (for example, by clicking on area 211), the arithmetic unit 11 controls the display device 151 to display image 220 (see Figure 4) showing the search results from the search unit 111. In image 220, the text data included in the prompt (in other words, the text data input to the large-scale language model) and the text data not included in the prompt may be visually distinguished.

[0026] When image 220 is displayed on the display device 151, the user can see the text data that has been input into the large-scale language model. For example, if the user recognizes that the text data that the user expects has not been input into the large-scale language model, the user may change the search score associated with each piece of text data included in the search results. Specifically, the user may instruct the large-scale language model to change the search score via the input device 14 so that the text data that the user desires is input into the large-scale language model (in other words, so that the text data that the user desires is included in the prompt). As a result, the modification unit 114 of the information processing device 10 may change the search scores included in the search results based on the user's instructions.

[0027] If the search score included in the search results is changed, the input unit 112 selects one or more text data items to include in the prompt from the text data included in the search results based on the changed search score. The input unit 112 sends a prompt containing the question and the selected one or more text data items to the server 20 via the communication device 13. As a result, the prompt is input to the large-scale language model.

[0028] If the search score is changed, the arithmetic unit 11 may associate one or more text data selected based on the changed search score with the question and store them in the storage device 12. Subsequently, if the user enters a new question via the input device 14, the search unit 111 may search the knowledge base 30 based on the new question. The search unit 111 may further determine whether a question similar to the new question is stored in the storage device 12. If it is determined that a question similar to the new question is stored in the storage device 12, the search unit 111 may adjust the search score included in the search results of the knowledge base 30 based on the new question, based on one or more text data associated with the question similar to the new question. For example, the search unit 111 may adjust the search score so that the one or more associated text data is included in the prompt.

[0029] (Technical effects) In this embodiment, a prompt containing one or more text data selected based on a search score modified by the user is input to the large-scale language model, thereby regenerating the large-scale language model's response to the user's question. In this case, since the text data desired by the user is input to the large-scale language model, the regenerated response can be expected to be the answer the user is looking for. Furthermore, in this embodiment, one or more text data selected based on the modified search score and the question are associated and stored in the storage device 12. When a new question is input, the search score included in the search results of the knowledge base 30 based on the new question may be adjusted based on one or more text data associated with question sentences similar to the new question. Therefore, the large-scale language model's response to the new question can be expected to be the answer the user is looking for. In this way, the information processing device 10 according to this embodiment can improve the response accuracy of the large-scale language model.

[0030] Various aspects of the invention derived from the embodiments described above are described below.

[0031] An information processing device according to one aspect of the invention comprises: a search means for searching a database based on a question entered by a user; an input means for inputting text data as part of the search results from the database obtained by the search means and the question into a large-scale language model; an acquisition means for obtaining an answer to the question generated by the large-scale language model; a display means for displaying the answer; and a receiving means capable of receiving user input, wherein when the receiving means receives user input indicating that the user selects the displayed answer, the display means displays the search results.

[0032] In the above-described embodiment, "Knowledge Base 30" corresponds to an example of a "database," "Search Unit 111" corresponds to an example of a "search means," "Input Unit 112" corresponds to an example of an "input means," "Acquisition Unit 113" corresponds to an example of an "acquisition means," "Input Device 14" corresponds to an example of a "reception means," and "Display Device 151" corresponds to an example of a "display means."

[0033] In the information processing device according to the above embodiment, the search results may include text data and a search score indicating the degree of relationship between the text data and the question.

[0034] The information processing device according to the above embodiment may include a modification means for modifying the search score when the receiving means receives input from the user indicating that the search score should be changed. In the above embodiment, the "modification unit 114" corresponds to an example of the "modification means".

[0035] In the information processing device according to the above embodiment, if the search score is changed, the input means may input one or more text data selected based on the changed search score and the question sentence into the large-scale language model.

[0036] The information processing device according to the above embodiment may include storage means for storing one or more text data selected based on the modified search score in association with the question sentence. In the above embodiment, "storage device 12" corresponds to an example of "storage means".

[0037] The present invention is not limited to the embodiments described above, and can be modified as appropriate without contradicting the gist or idea of ​​the invention as can be read from the claims and specification as a whole. Information processing devices that involve such modifications are also included within the technical scope of the present invention. [Explanation of Symbols]

[0038] 1... Information processing system, 10... Information processing device, 20... Server, 30... Knowledge base, 111... Search unit, 112... Input unit, 113... Acquisition unit, 114... Change unit

Claims

1. A search method that searches a database based on a question entered by the user, An input means for inputting text data as part of the search results of the database obtained by the search means and the question text into a large-scale language model, An acquisition means for obtaining the answer to the question sentence generated by the aforementioned large-scale language model, A display means for displaying the aforementioned answer, A receiving means capable of receiving user input, Equipped with, When the receiving means receives input from the user indicating that they have selected the displayed answer, the display means displays the search results. Information processing device.

2. The search results include text data and a search score indicating the degree of relevance between the text data and the question. The information processing apparatus according to claim 1.

3. If the receiving means receives input from the user indicating that the search score should be changed, it includes a means for changing the search score. The information processing apparatus according to claim 2.

4. If the search score is changed, the input means inputs one or more text data selected based on the changed search score and the question sentence into the large-scale language model. The information processing apparatus according to claim 3.

5. The system includes a storage means for associating and storing one or more text data selected based on the modified search score with the question text. The information processing apparatus according to claim 4.

Images