Information processing apparatus, control method, and storage medium

Abstract

To provide an information processing apparatus, a control method, and a program capable of suitably supporting reading by a user.SOLUTION: An acquiring unit that acquires a selected text selected by a user from among texts displayed to the user, a determining unit that determines, based on the selected text, information related to an intention of the user selecting the selected text, a deciding unit that decides a question sentence creation plan based on the information related to the intention and the selected text, and a generating unit that generates, based on the question sentence creation plan, a prompt for instructing a generative AI model to generate the question sentence.SELECTED DRAWING: Figure 2

Description

[Technical Field]

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

[0002] Some users of e-books have had trouble with kanji characters they don't know how to read, unfamiliar terms and words, or difficult sentences they can't understand. To help these users, e-books offer functions such as a search function that allows you to select a word and search for it.

[0003] Patent Document 1 also discloses a technology for searching a document displayed on a mobile device, where a system presents words taking into account the context of a selected word. This technology presents words containing the selected word and presents sentences that conform to a set syntax. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Special Publication No. 2017-525026 Summary of the Invention [Problem to be solved by the invention]

[0005] As such, conventional technologies merely search for words or present sentences that conform to a set syntax, and do not adequately support the reading of users of e-books.

[0006] One aspect of the present disclosure provides an information processing device, a control method, and a program that can favorably support a user's reading. [Means for solving the problem]

[0007] An information processing device according to one embodiment of the present disclosure includes an acquisition unit that acquires selected text selected by a user from text displayed to the user; a determination unit that determines information regarding the intention of the user in selecting the selected text based on the selected text; a determination unit that determines a policy for creating a question sentence based on the information regarding the intention and the selected text; and a generation unit that generates a prompt that instructs a generative AI model to generate the question sentence based on the policy for creating the question sentence. [Brief explanation of the drawings]

[0008] [Figure 1] FIG. 1 illustrates an example of a reading support system. [Figure 2] FIG. 10 is a diagram illustrating a display example of an information processing device. [Figure 3] FIG. 2 illustrates an example of the configuration of a knowledge server. [Figure 4A] FIG. 10 is a diagram illustrating an example of the configuration of supplemental question information. [Figure 4B] FIG. 10 is a diagram showing an example of the configuration of content question information. [Figure 4C] FIG. 10 is a diagram illustrating an example of the configuration of question tag information. [Figure 4D] FIG. 10 is a diagram illustrating an example of the configuration of tag probability information. [Figure 4E] FIG. 10 is a diagram illustrating an example of the configuration of user attribute information. [Figure 4F] FIG. 10 is a diagram illustrating an example of the configuration of reference evaluation information. [Figure 5] FIG. 1 illustrates an example of the configuration of an information processing device. [Figure 6] FIG. 1 illustrates an example of the configuration of a generating device. [Figure 7] FIG. 2 is a diagram illustrating an example of the configuration of a tag question response server. [Figure 8] 10 is a flowchart showing the flow of a support process. [Figure 9] 10 is a flowchart showing the flow of an intention determination process. [Figure 10] 10 is a flowchart showing the flow of a creation policy determination process. [Figure 11]FIG. 10 is a diagram illustrating an example of a prompt. [Figure 12] FIG. 2 is a diagram illustrating an example of the hardware configuration of an information processing device, a knowledge server, a generation device, and a tag question response server. DETAILED DESCRIPTION OF THE INVENTION

[0009] Hereinafter, an embodiment according to one aspect of the present disclosure will be described with reference to the drawings. Note that the embodiment described below is an example, and the embodiment to which the present disclosure is applied is not limited to the following embodiment.

[0010] 1 is a diagram illustrating an example of a reading support system 10 according to an embodiment of the present disclosure. The reading support system 10 includes an information processing device 100, a knowledge server 200, a generation device 300, and a tag question response server 400. Each device is connected to a network NW, and the information processing device 100 can communicate with each of the knowledge server 200, the generation device 300, and the tag question response server 400.

[0011] The information processing device 100 is, for example, a smartphone or a tablet. The information processing device 100 includes a display device that displays an e-book as text to a user using the information processing device 100 and also displays images included in the e-book. The information processing device 100 acquires selected text selected by the user from the text displayed to the user, generates a prompt, and transmits the prompt to the generation device 300. The information processing device 100 displays a question and an answer generated by the generation device 300 in response to the prompt. Here, a prompt refers to an instruction or question input by a user in a dialogue with a generative AI model such as a large-scale language model (LLM). In the following description, the above-described processing by the information processing device 100 (processing of acquiring selected text, generating a prompt, and displaying a question and an answer) may be referred to as "assistance processing."

[0012] The knowledge server 200 stores various information required for the information processing device 100 to generate a prompt. The generation device 300 generates an answer corresponding to the prompt sent from the information processing device 100 and sends it to the information processing device 100. The tag question response server 400 outputs the tag output by the information processing device 100 when generating the prompt, and a question sentence that is highly related to the tag.

[0013] A specific example of the display of the information processing device 100 will be described with reference to the drawings. Fig. 2 is a diagram showing an example of the display of the information processing device 100. Fig. 2 shows a screen 110 showing a state in which text has been selected, and a screen 120 showing a state after the assistance process has been executed.

[0014] Screen 110 is a screen that shows the state in which text is being displayed to the user. Note that in FIG. 2, parts of the text are indicated by dots. Selected text 111 displayed on screen 110 indicates the text selected by the user from the text displayed on screen 110. Also displayed on screen 110 are question buttons 112. Question buttons 112 include two buttons: "Look up in dictionary" and "Ask AI." Of these, "Look up in dictionary" is a button for looking up selected text 111 in a dictionary. "Ask AI" is a button for executing the above-mentioned support processing.

[0015] The screen 120 displays a question field 121, an answer field 122, and a reference button 123. The question field 121 displays a question generated by the generation device 300. The answer field 122 displays an answer generated by the generation device 300. In the case of the screen 120, three answers are displayed for one question. The reference button 123 is a button that the user taps if the question and answer are helpful. Note that, although only one question is shown on the screen 120 as an example, there may be multiple questions.

[0016] In this way, when a user encounters a problem while reading, the information processing device 100 can provide assistance using the generative AI model by selecting the relevant text. This eliminates the need for the user to switch applications to a dictionary or search browser, allowing the user to solve the problem in this application without interrupting the application displaying content such as an e-book.

[0017] Next, the knowledge server 200 will be described. Fig. 3 is a diagram showing an example of the configuration of the knowledge server 200. In Fig. 3, the knowledge server 200 includes a storage unit 202. The storage unit 202 stores supplemental question information 211, content question information 212, question tag information 213, tag probability information 214, user attribute information 215, and reference evaluation information 216.

[0018] These various types of information will be explained below. In the following explanation, duplicate explanations of data included in each piece of information that have already been mentioned will be omitted.

[0019] 4A is a diagram showing an example of the configuration of the supplemental question information 211. The supplemental question information 211 includes words included in the content, question content indicating what kind of question is being asked about the words, and the probability that the question is asked with that question content. Note that the content in this embodiment includes text displayed by the information processing device 100, such as an e-book or a web page, and may also include images and videos.

[0020] 4B is a diagram showing an example of the configuration of content question information 212. Content question information 212 includes content identification information and a question. The content identification information is information for uniquely identifying content. Content question information 212 is a question posed to content identified by the content identification information.

[0021] FIG. 4C is a diagram showing an example of the configuration of question tag information 213. Question tag information 213 is composed of a question sentence and one or more tags. Here, a tag is a noun included in the question sentence. For example, if the question sentence is "Please tell me about the characteristics of cash flow management at XX Co., Ltd.", the tags would be "XX Co., Ltd.", "cash flow management," etc. Note that proper nouns such as "XX Co., Ltd.", nouns with well-known meanings such as "characteristics," numerals, formal nouns, and pronouns may be deleted from the tags. In the following explanation, "tag" and "noun" are considered to be the same thing.

[0022] FIG. 4D is a diagram showing an example of the configuration of tag probability information 214. Tag probability information 214 includes tags and probabilities, and indicates the probability that a user will ask a question about a tag. FIG. 4E is a diagram showing an example of the configuration of user attribute information 215. User attribute information 215 includes age, gender, status, and favorite genre. Age indicates the user's age, and gender indicates the user's gender. Status indicates the user's status (working adult, student, etc.). Favorite genre indicates the user's favorite book genre.

[0023] FIG. 4F is a diagram showing an example of the configuration of reference rating information 216. Reference rating information 216 includes content identification information, a question, and a reference ratio. This information is for storing the ratio (reference ratio) at which reference button 123 described in FIG. 2 is tapped. As shown in FIG. 2, information processing device 100 displays to the user a question and its answer for content identified by the content identification information. At this time, the question that is the subject of evaluation by the user is the question in the reference rating information, and the reference ratio indicates the value obtained by dividing the number of times reference button 123 is tapped by the total number of times the question is displayed.

[0024] Next, a description will be given of an example configuration of the information processing device 100. Fig. 5 is a diagram showing an example configuration of the information processing device 100. In Fig. 5, the information processing device 100 is made up of a display unit 101, an operation unit 102, an acquisition unit 103, a determination unit 104, a determination unit 105, a generation unit 106, a prompt transmission unit 107, and a response reception unit 108.

[0025] The display unit 101 displays various information (text, images, videos, etc.) on a display device provided in the information processing device 100. The operation unit 102 is a hard key or touch panel operated by the user. The acquisition unit 103 acquires selected text selected by the user from the text displayed to the user. The determination unit 104 determines information related to the intention of the user in selecting the selected text based on the selected text. The method of determining information related to the intention will be described in detail later.

[0026] The determination unit 105 determines a question creation policy based on the information related to the determined intention and the selected text. The method for determining the creation policy will be described in detail later. The generation unit 106 uses Retrieval-Augmented Generation (RAG), a type of prompt augmentation technology, to generate a prompt that instructs the generation AI model of the generation device 300 to generate a question based on the question creation policy. A common application example of RAG is when a user makes a request to an LLM by searching for similar documents in a search system in advance and requesting those documents from the LLM. This allows the LLM to generate an answer based on specific documents, such as company information.

[0027] The prompt sending unit 107 sends the prompt generated by the generation unit 106 to the generation device 300. The response receiving unit 108 receives a response generated by the generation device 300 in response to the prompt. The received response is displayed on the display device by the display unit 101, as shown in FIG.

[0028] Next, a description will be given of the generation device 300. Fig. 6 is a diagram showing an example of the configuration of the generation device 300. The generation device 300 includes a prompt receiving unit 301, a generation AI model 302, and a response sending unit 303.

[0029] The prompt receiving unit 301 receives a prompt sent by the information processing device 100 and inputs it to the generative AI model 302. The generative AI model 302 functions as a generative AI as described above, generates a response corresponding to the prompt, and outputs it to the response sending unit 303. The response sending unit 303 sends the response output by the generative AI model 302 to the information processing device 100.

[0030] Next, the tag-question response server 400 will be described. FIG. 7 is a diagram showing an example of the configuration of the tag-question response server 400. The tag-question response server 400 includes a tag receiving unit 401, a trained model 402, and a question sending unit 403. The tag receiving unit 401 receives tags sent from the information processing device 100 and inputs them to the trained model 402. The trained model 402 is a trained model that has been trained to learn questions that are highly related to tags. As a result, the trained model 402 outputs questions that are highly related to the tags output by the tag receiving unit 401 to the question sending unit 403. The question sending unit 403 sends the questions output by the trained model 402 to the information processing device 100.

[0031] Note that a GNN (Graph Neural Network), for example, may be used as the trained model 402. In this case, a tag node and a question node may be generated, and training may be performed by treating a case where a user has asked a question about a tag as a positive example, and a case where a user has not asked a question about a tag as a negative example.

[0032] Based on the above-described configuration, the flow of the support process will be explained using a flowchart. Fig. 8 is a flowchart showing the flow of the support process. The flowchart shown in Fig. 8 shows the process that starts when the user selects text and taps "Ask AI" shown in Fig. 2.

[0033] 8, the acquisition unit 103 of the information processing device 100 acquires selected text selected by a user (step S101). The selected text is acquired by a system call of an OS (Operating System), etc. The determination unit 104 executes an intention determination process to determine information about the intention of the user who selected the selected text, based on the selected text (step S102). The intention determination process will be described in detail later. The determination unit 105 executes a creation policy determination process to determine a question creation policy, based on the information about the intention and the selected text (step S103). The creation policy determination process will be described in detail later. In this creation policy determination process, nouns (tags) are extracted from the selected text.

[0034] The generation unit 106 searches the question supplement information 211 for question contents corresponding to the tags extracted in the creation policy determination process, acquires question contents that are asked with a predetermined probability or higher, and supplements the question (step S104). Next, the generation unit 106 searches for words within text included in a predetermined range based on the selected text (for example, within the page of the e-book on which the selected text is written, or within a predetermined number of characters before and after the selected text), and acquires words with a high occurrence frequency (for example, the most frequent word, the top three most frequent words, or words with an occurrence frequency greater than a predetermined value) (step S105).

[0035] Based on the question creation policy, a prompt is generated to instruct the generation AI model to generate a question (step S106). At this time, the generation unit 106 generates the prompt using the user's attribute information by referring to the question acquired in step S104, the words acquired in step S105, and the user attribute information.

[0036] The prompt sending unit 107 sends the prompt generated by the generation unit 106 (step S107), and the response receiving unit 108 receives the response sent by the generation device 300 (step S108). The display unit 101 displays the question and answer indicated in the response on the display device, as shown in Fig. 2 (step S109). The display unit 101 updates the reference evaluation information depending on whether the reference button 123 has been tapped (step S110).

[0037] Next, the intention determination process of step S102 by the determination unit 104 will be described. In this embodiment, information regarding the user's intention is classified into three types (P, Q, R). P indicates an intention to know the meaning of a word when the selected text consists of only one word. Q indicates an intention to know how to read the word when the selected text consists of only one word. R indicates an intention to know the meaning of the selected text (sentence) (for example, its meaning in a book) when the selected text consists of multiple words, or indicates an intention other than P and Q.

[0038] 9 is a flowchart showing the flow of the intention determination process. In FIG. 9, the determination unit 104 determines whether the selected text is only one word (step S201). If the selected text is only one word (step S201: YES), the determination unit 104 determines the information regarding the intention as P or Q (step S202) and ends the process. On the other hand, if the selected text is not only one word (step S201: NO), the determination unit 104 determines the information regarding the intention as R (step S203) and ends the process.

[0039] Next, the creation policy determination process of step S103 by the determination unit 105 will be described. Fig. 10 is a flowchart showing the flow of the creation policy determination process. In Fig. 10, the determination unit 105 determines whether the information on the intention determined by the determination unit 104 is P or Q (step S301). If the information on the intention is P or Q (step S301: YES), the determination unit 105 determines the question sentence creation policy to be a policy of asking about the meaning or pronunciation of a word (step S302), and ends the process.

[0040] If the information on the intention is R (step S301: NO), the determination unit 105 performs morphological analysis on the selected text and extracts nouns (step S303). The determination unit 105 obtains the probability for each extracted tag using the tag probability information 214 (step S304). The determination unit 105 obtains tags whose probabilities are equal to or greater than a predetermined threshold (step S305). Here, only the tag with the highest probability may be obtained.

[0041] The determining unit 105 transmits each acquired tag to the tag question server 400 and acquires questions that are highly related to each tag (step S306). At this time, the determining unit 105 may acquire questions with a high reference ratio from the reference evaluation information (for example, the question with the highest ratio, the top three questions with reference ratios, or questions with reference ratios greater than a predetermined value). The determining unit 105 decides to use the acquired questions as reference questions (step S307) and terminates the process.

[0042] FIG. 11 is a diagram showing an example of a prompt generated by the generation unit 106. As shown in FIG. 11, the prompt 500 includes a "task," "input information," "condition," and "output format." The generation unit 106 holds the "task" and "output format" as fixed phrases and sets the "input information" and "condition." As shown in the "output format," the prompt 500 also includes an instruction to generate an answer corresponding to the question. The "input information" includes a selection text field 501 and an attribute field 502. The "condition" includes a reference text field 503 and a frequency field 504.

[0043] The selected text field 501 is set with the selected text acquired by the acquisition unit 103. The attribute field 502 indicates an instruction for generating a question according to the user's attributes. The attribute field 502 is set with the age, gender, status, and favorite genre indicated in the user attribute information 215. The reference field 503 is set with the question received from the tag question response server 400 and the probability (%) corresponding to the question obtained from the question probability information. In the case of FIG. 11, three questions and their corresponding probabilities are set. Thus, the prompt 500 includes an instruction for generating a question based on the reference question to be used as a reference when generating a question and the probability that the reference question will be asked by the user. The frequency field 504 is set with the words acquired in step S105 and their frequencies.

[0044] <Effects> Generating the above-described prompt 500 can provide optimal support for the user's reading. Specifically, since the prompt 500 includes attribute information, it is possible to increase the likelihood that a question that is more suited to the user's position will be generated. Furthermore, since the prompt 500 includes reference sentences and probabilities, it is possible to increase the likelihood that a question that many users want to ask will be generated. At this time, the reference sentences and probabilities are provided to the generative AI model 302, so it is also possible to train the generative AI model 302. Furthermore, since the prompt 500 includes frequently occurring words and their frequencies that are considered to be highly relevant to the selected text, it is possible to increase the likelihood that a question that is more in line with the contents of the book will be generated.

[0045] In the above-described embodiment, an e-book displayed on the information processing device 100 has been described as an example of a reading object. However, the present embodiment is not limited to this, and may be applied to any text displayed on the information processing device 100, such as text on a web page or text on a social networking site. The object selected by the user may be something that can be converted into text by applying character recognition technology, such as characters embedded in an image. Furthermore, the object that the user can select may be an image or video in addition to text. When an image or video is selected, the information processing device 100 generates a prompt in the generative AI model to generate a question or answer, assuming that the image or video has been selected by the user.

[0046] <Variation 1> In the intention determination process of step S102, an AI model may be used to determine information about the intention (either P, Q, or R). Specifically, when a question about the meaning of the selected text is asked, P may be learned as the correct answer, when a question about the pronunciation of the selected text is asked, Q may be learned as the correct answer, and when neither P nor Q is asked, R may be learned as the correct answer. This increases the likelihood that a question that matches the user's intention will be generated, thereby providing optimal support for the user's reading. Note that, for a certain selected text, if the number of times P was obtained is less than a predetermined value, or if the value obtained by dividing the number of times P was obtained by the number of times R was obtained is less than a predetermined value, the information about the intention corresponding to the selected text may always be determined to be R without using an AI model.

[0047] <Variation 2> In the process of supplementing questions in step S104, an AI model may be used to supplement the questions. For example, an AI model may be prepared that indicates the relevance of each question (the relevance of the same user asking a question after a search has been performed once), and the questions may be supplemented using this AI model. Specifically, if a user who asked the question "The meaning of cash flow management" asks "Why is cash flow management important?", the AI ​​model may be trained to recognize that there is a relevance. When using the GNN described above, for example, the distance between question nodes corresponding to questions that are relevant as positive examples is shortened.

[0048] <Variation 3> In the above step S306, the determining unit 105 transmits each acquired tag to the tag question server 400 and acquires a question sentence highly relevant to each tag, but the determining unit 105 may estimate a question sentence based on information about the user (for example, the user's reading history data) and set this question sentence as a highly relevant question sentence. As an estimation method, for example, there is a method of determining a question sentence previously asked by a user who has similar reading history data as the relevant user as a highly relevant question sentence.

[0049] <Variation 4> A means for the user to ask a follow-up question in response to the answer sentence may be provided. The information processing device 100 may present information (which may be other than text) indicating candidate follow-up questions. Candidate follow-up questions may be stored as follow-up questions that have been frequently asked after a question in the past.

[0050] <Variation 5> The answer sent by the generating device 300 as a response may be composed of not only text but also any one or a combination of images, audio, video, etc., and the user may be able to select the type of composition.

[0051] <Variation 6> 11 causes a question sentence and an answer sentence to be output together, but it may also be a prompt that instructs to output only the question sentence. In this case, generation unit 106 may generate a prompt that instructs to output multiple question sentences and to sort and output the question sentences in order of the likelihood that they are intended by the user.

[0052] The information processing device 100 displays to the user a plurality of questions received as responses from the generation device 300. The information processing device 100 then prompts the user to select a question from the plurality of questions that most closely matches the user's intended question. The selected question may be used, for example, as supplemental question information or content question information. The tag question response server 400 may also be trained using tags obtained from the selected text and the selected question. Furthermore, in the intention determination process, an AI model may be used to determine information regarding the intention (P, Q, or R). Specifically, if the selected question asks about the meaning of a word, P may be learned as the correct answer. If the selected question asks about the pronunciation of a word, Q may be learned as the correct answer. If neither P nor Q applies, R may be learned as the correct answer. This increases the likelihood that a question more closely matches the user's intended question, thereby providing effective support for the user's reading.

[0053] <Variation 7> Among the above-mentioned information about intention, R can be further subdivided as follows: - The desire to learn about the cultural background. This is because some books contain expressions that cannot be understood without knowledge of the cultural background and historical facts. The user does not understand grammatical rules or structures and wants to know them. This is especially true when the user is reading a book written in a language other than their native language, where grammatical rules or structures may not be understandable. The desire to understand the author's intentions. This is because you may not understand what the author is trying to convey or what a particular expression or metaphor refers to. The desire to learn about story developments, as the story development and character motivations may not be clear. The desire to learn specialized knowledge. This is because specialized knowledge or theory in a particular field may be required but is sometimes lacking. By further subdividing the information about the intention in this way, it is possible to determine a creation policy that is more in line with the intention, thereby increasing the likelihood that a question that is more in line with the user's intention will be generated.

[0054] <Variation 8> In the system configuration of the reading assistance system 10 described above, the generative AI model 302 is implemented in a device connected to the information processing device 100 via a different network, but it may also be implemented in the information processing device 100. In addition, the prompt generation function using RAG implemented in the information processing device 100 may be implemented together with the generative AI model in a device connected to the network.

[0055] Furthermore, the functional units of the information processing device 100, the knowledge server 200, the generating device 300, and the tag question response server 400 may be distributed and implemented on the cloud. Also, each functional unit may be implemented in multiple information processing devices. Furthermore, the same functional unit may be realized by multiple information processing devices.

[0056] <Hardware configuration example> The information processing device 100, the knowledge server 200, the generation device 300, the tag question response server 400, and the like according to an embodiment of the present disclosure may function as computers that perform the processing of the present disclosure. Fig. 12 is a diagram illustrating an example of the hardware configuration of the information processing device 100, the knowledge server 200, the generation device 300, and the tag question response server 400 according to an embodiment of the present disclosure. The information processing device 100, the knowledge server 200, the generation device 300, and the tag question response server 400 described above may be physically configured as a computer device including a processor 1001, a memory 1002, a storage 1003, a communication device 1004, an input device 1005, an output device 1006, a bus 1007, and the like.

[0057] In the following description, the term "apparatus" can be interpreted as a circuit, a device, a unit, etc. The hardware configurations of the information processing device 100, the knowledge server 200, and the generation device 300 may be configured to include one or more of the devices shown in the drawings, or may be configured to exclude some of the devices.

[0058] Each function of the information processing device 100, knowledge server 200, generation device 300, and tag question response server 400 is realized by loading predetermined software (programs) onto hardware such as a processor 1001 and memory 1002, causing the processor 1001 to perform calculations, control communication via the communication device 1004, and control at least one of reading and writing data in the memory 1002 and storage 1003.

[0059] The processor 1001 controls the entire computer by running, for example, an operating system, and may be configured as a central processing unit (CPU) including an interface with peripheral devices, a control device, an arithmetic unit, a register, etc.

[0060] Furthermore, the processor 1001 reads programs (program codes), software modules, data, etc. from at least one of the storage 1003 and the communication device 1004 into the memory 1002 and executes various processes in accordance with these. The programs used are those that cause a computer to execute at least some of the operations described in the above-described embodiments. For example, each unit of the information processing device 100 may be implemented by a control program stored in the memory 1002 and running on the processor 1001, and similar implementations may be made for other functional blocks. While the above-described various processes have been described as being executed by one processor 1001, they may also be executed simultaneously or sequentially by two or more processors 1001. The processor 1001 may be implemented by one or more chips. The programs may also be transmitted from a network via a telecommunications line.

[0061] The memory 1002 is a computer-readable recording medium and may be configured, for example, by at least one of a read-only memory (ROM), an erasable programmable ROM (EPROM), an electrically erasable programmable ROM (EEPROM), a random access memory (RAM), etc. The memory 1002 may also be called a register, a cache, a main memory (primary storage device), etc. The memory 1002 can store executable programs (program codes), software modules, etc. for implementing a wireless communication method according to an embodiment of the present disclosure.

[0062] Storage 1003 is a computer-readable recording medium, and may be composed of at least one of, for example, an optical disk such as a CD-ROM (Compact Disc ROM), a hard disk drive, a flexible disk, a magneto-optical disk (e.g., a compact disk, a digital versatile disk, a Blue-ray® disk), a smart card, a flash memory (e.g., a card, a stick, a key drive), a floppy disk, a magnetic strip, etc. Storage 1003 may also be referred to as an auxiliary storage device. The above-mentioned storage medium may be, for example, a database, a server, or other appropriate medium including at least one of memory 1002 and storage 1003.

[0063] The communication device 1004 is hardware (transmission / reception device) for communicating between computers via at least one of a wired network and a wireless network, and is also called, for example, a network device, a network controller, a network card, or a communication module.

[0064] The input device 1005 is an input device (for example, a keyboard, a mouse, a microphone, a switch, a button, a sensor, etc.) that receives input from the outside. The output device 1006 is an output device (for example, a display, a speaker, an LED lamp, etc.) that outputs to the outside. The input device 1005 and the output device 1006 may be integrated into one device (for example, a touch panel).

[0065] Furthermore, each device, such as the processor 1001 and the memory 1002, is connected by a bus 1007 for communicating information. The bus 1007 may be configured using a single bus, or may be configured using different buses between each device.

[0066] Furthermore, the information processing device 100, the knowledge server 200, and the generating device 300 may be configured to include hardware such as a microprocessor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a programmable logic device (PLD), or a field programmable gate array (FPGA), and some or all of the functional blocks may be realized by the hardware. For example, the processor 1001 may be implemented using at least one of these pieces of hardware. [Industrial Applicability]

[0067] One aspect of the present disclosure is useful for assisting in reading electronic books. [Explanation of symbols]

[0068] 10 Reading Support System 100 Information processing device 103 Acquisition Department 104 Judgment section 105 Decision Section 200 Knowledge Server 300 generator 302 Generative AI Models 400 Tag Question Response Server 1001 processor 1002 memory 1003 Storage 1004 Communication equipment 1005 Input Device 1006 Output Device 1007 Bus

Claims

1. an acquisition unit that acquires selected text selected by the user from among the texts displayed to the user; a determination unit that determines information regarding the user's intention in selecting the selected text based on the selected text; a determination unit that determines a question creation policy based on the information about the intention and the selected text; A generation unit that generates a prompt that instructs a generative AI model to generate the question sentence based on the question sentence creation policy; An information processing device comprising:

2. 2. The information processing device according to claim 1, wherein the determination unit determines that the information regarding the intention indicates an intention to know the meaning or pronunciation of a word when the selected text consists of only one word, and determines that the information regarding the intention indicates an intention to know the meaning of the selected text when the selected text consists of multiple words.

3. 2. The information processing device of claim 1, wherein the determination unit determines the creation policy to be a policy of asking about the meaning or pronunciation of a word when the information regarding the intention determined by the determination unit indicates an intention to know the meaning or pronunciation of a word, and when the information regarding the intention determined by the determination unit indicates an intention to know the meaning or pronunciation of a word, the determination unit determines the creation policy to be a policy of performing morphological analysis on the selected text, extracting nouns, obtaining a question sentence that is highly related to the extracted nouns, and using the obtained question sentence as a reference question sentence.

4. The information processing device according to claim 1 , wherein the generating unit generates the prompt including an instruction to generate an answer sentence corresponding to the question sentence.

5. The information processing device according to claim 1 , wherein the generating unit generates the prompt including an instruction to generate the question sentence according to an attribute of the user.

6. The information processing device according to claim 1 , wherein the generation unit generates the prompt including an instruction to generate the question based on a reference question sentence to be used as a reference when generating the question sentence and a probability that the reference question sentence will be asked by the user.

7. The information processing device according to claim 1 , wherein the generating unit generates the prompt including an instruction to generate the question sentence based on nouns in text included in a predetermined range based on the selected text and the frequency of appearance of the nouns.

8. The information processing device according to claim 4 , further comprising a display unit that displays the question sentence and the answer sentence corresponding to the question sentence generated by the generative AI model.

9. The information processing device Obtaining a selected text from the text displayed to the user, selected by the user; determining information regarding an intent behind the user selecting the selected text based on the selected text; determining a question generation policy based on the information about the intention and the selected text; generating a prompt that instructs a generative AI model to generate the question sentence based on the question sentence generation policy; Control method.

10. On the computer, Obtaining a selected text from the text displayed to the user, selected by the user; determining information regarding an intent behind the user selecting the selected text based on the selected text; determining a question generation policy based on the information about the intention and the selected text; generating a prompt that instructs a generative AI model to generate the question sentence based on the question sentence generation policy; A program for executing a process.

Images