Support apparatus, support system, support method, and program

The assistance device tailors document conversion to individual characteristics, addressing reading difficulties beyond vocabulary and grammar, thereby improving comprehension for diverse readers.

JP2025128523APending Publication Date: 2025-09-03NEC CORP
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
JP2024025227
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Filing Date
2024-02-22
Publication Date
2025-09-03

AI Technical Summary

Technical Problem

Existing reading assistance systems are ineffective for readers whose comprehension difficulties stem from factors other than vocabulary or grammatical ability, necessitating a system that provides tailored support based on individual characteristics.

Method used

An assistance device and system that acquires recipient information, calculates an evaluation value, generates prompts with constraints, and uses a generative model to convert documents for easier reading, considering factors like vocabulary, grammatical ability, and individual comprehension strengths.

Benefits of technology

Provides personalized reading assistance by converting documents to suit the reader's characteristics, enhancing comprehension for individuals with diverse reading challenges.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2025128523000001_ABST
    Figure 2025128523000001_ABST
Patent Text Reader

Abstract

To provide a support apparatus capable of performing reading support according to characteristics of a support target person.SOLUTION: A support apparatus comprises: a support target person information acquisition unit that acquires support target person information representing characteristics of a support target person who is a target person of reading support; a support evaluation value calculation unit that calculates a support evaluation value indicating a degree of reading support for the support target person according to the support target person information; a target document acquisition unit that acquires a target document that is the target of reading support; a prompt generation unit that generates a prompt including constraint conditions according to the support evaluation value and the target document; a transformed document generation unit that generates output information output from a learned generative model in response to input of the prompt as a transformed document; and an output unit that outputs the generated transformed document.SELECTED DRAWING: Figure 9
Need to check novelty before this filing date? Find Prior Art

Description

[Technical Field]

[0001] The present disclosure relates to an assistance device, an assistance system, an assistance method, and a program. [Background technology]

[0002] There are known techniques for assisting readers in reading comprehension. Patent Document 1 discloses an assist system that infers a user's level of understanding of content from the movement of the user's viewpoint and provides assist information according to the inferred level of understanding. When the content is a sentence, the assist information is information indicating the meaning of words contained in the sentence, the grammar of the sentence, etc. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] International Publication No. 2022 / 070747 Summary of the Invention [Problem to be solved by the invention]

[0004] Factors that impede reading comprehension vary depending on the characteristics of the reader. When factors that impede reading comprehension are due to characteristics such as the reader's vocabulary or grammatical ability, the assistance system described in Patent Document 1 can be used to effectively support the reader's reading comprehension. However, when factors that impede reading comprehension are due to characteristics other than the reader's vocabulary or grammatical ability, the assistance system described in Patent Document 1 is less effective in supporting the reader's reading comprehension. Therefore, it is desirable to build a system that can provide reading comprehension assistance to readers that is tailored to the characteristics of the reader who is the target of support.

[0005] An object of the present disclosure is to provide an assistance device, an assistance system, an assistance method, and a program that are capable of providing reading assistance according to the characteristics of a person receiving assistance. [Means for solving the problem]

[0006] An assistance device according to one aspect of the present disclosure includes an assistance recipient information acquisition unit that acquires assistance recipient information that represents the characteristics of the assistance recipient who is the target of reading assistance; an assistance evaluation value calculation unit that calculates an assistance evaluation value that indicates the degree of reading assistance for the assistance recipient according to the assistance recipient information; a target document acquisition unit that acquires a target document that is the target of reading assistance; a prompt generation unit that generates a prompt that includes a constraint condition according to the assistance evaluation value and the target document; a converted document generation unit that generates output information output from a trained generative model according to input of the prompt as a converted document; and an output unit that outputs the generated converted document.

[0007] In one aspect of the assistance method of the present disclosure, a computer acquires assistance recipient information that represents the characteristics of the assistance recipient who is the target of reading assistance, calculates an assistance evaluation value that indicates the degree of reading assistance for the assistance recipient based on the assistance recipient information, acquires a target document that is the target of reading assistance, generates a prompt that includes a constraint based on the assistance evaluation value and the target document, generates output information that is output from a trained generative model based on input of the prompt as a converted document, and outputs the generated converted document.

[0008] A program according to one aspect of the present disclosure causes a computer to perform the following processes: acquire support recipient information that represents the characteristics of the support recipient who is the target of reading support; calculate a support evaluation value that indicates the degree of reading support for the support recipient based on the support recipient information; acquire a target document that is the target of reading support; generate a prompt that includes a constraint based on the support evaluation value and the target document; generate output information that is output from a trained generative model in response to input of the prompt as a converted document; and output the generated converted document. [Effects of the Invention]

[0009] According to the present disclosure, it is possible to provide an assistance device, an assistance system, an assistance method, and a program that are capable of providing reading assistance according to the characteristics of the assistance recipient. [Brief explanation of the drawings]

[0010] [Figure 1] 1 is a conceptual diagram for explaining an overview of a support system according to the present disclosure. [Figure 2] 1 is a block diagram illustrating an example of a configuration of a support system according to the present disclosure. [Figure 3] FIG. 10 is a diagram illustrating an example of a calculation table in the present disclosure. [Figure 4] FIG. 10 is a diagram illustrating an example of a conversion instruction reception screen according to the present disclosure. [Figure 5] FIG. 10 is a diagram illustrating an example of a constraint condition determination table in the present disclosure. [Figure 6] 10 is an example of a conversion instruction reception screen in the present disclosure. [Figure 7] FIG. 2 is a diagram illustrating an example of a prompt generated by a prompt generation unit according to the present disclosure. [Figure 8] 10 is a flowchart illustrating an example of an operation of the assistance device according to the present disclosure. [Figure 9] 1 is a block diagram illustrating an example of a configuration of a support device according to the present disclosure. [Figure 10] 10 is a flowchart illustrating an example of an operation of the assistance device according to the present disclosure. [Figure 11] FIG. 2 is a block diagram showing an example of a hardware configuration for executing control and processing in each embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0011] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. The embodiments described below are limited in a manner that is technically preferable for implementing the present disclosure, but the scope of the invention is not limited to the following. In all drawings used to describe the following embodiments, the same reference numerals are used for similar parts unless otherwise stated. In the following embodiments, repeated explanations of similar configurations and operations may be omitted. Furthermore, the GUIs (Graphical User Interfaces) shown in the drawings are merely examples and may be replaced with other GUIs.

[0012] In this disclosure, the term "support recipient" refers to a reader who is the target of reading support. Also, in this disclosure, the term "target document" refers to a document that is the target of reading support.

[0013] (First embodiment) First, a support system according to a first embodiment will be described with reference to the drawings. The support system of this embodiment converts a document for a support recipient so that the support recipient can easily read and understand it. The support system of this embodiment provides the converted document to the support recipient.

[0014] (overview) FIG. 1 is a conceptual diagram for explaining an overview of the assistance system according to the present disclosure. An overview of the assistance system 1 will be explained using FIG. 1. The assistance system 1 converts a target document to be read and understood by the assistance recipient into a converted document in accordance with the characteristics of the assistance recipient. The converted document is a document expressed in a format that is easy to read and understand for the characteristics of the assistance recipient. For example, the converted document may be expressed in a format that combines a picture or audio with the document.

[0015] The assistance device 10 acquires assistance recipient information, which is information that represents the characteristics of the assistance recipient. For example, the assistance recipient information is information that represents the characteristics unique to the assistance recipient, such as skills and personality. The assistance device 10 uses the assistance recipient information to calculate an assistance evaluation value for the assistance recipient. The assistance evaluation value is a numerical value that represents the degree of assistance required for the assistance recipient to read the target document. The assistance device 10 also acquires a target document that is the target of reading assistance.

[0016] The assistance device 10 generates a prompt (instruction sentence) using the assistance evaluation value and constraints determined in advance according to the assistance evaluation value. The prompt includes an instruction to output desired output information from the generative model. The prompt also includes text information of the target document. The generative model is a trained model that has been trained using a huge amount of data. The generative model generates output information in accordance with the input prompt. The assistance device 10 inputs the generated prompt into the generative model. The generative model generates a converted document that has been converted in accordance with the prompt as output information. The assistance device 10 outputs the generated converted document. The converted document has content that is easy for the assistance recipient to read and understand. Therefore, the assistance recipient can understand the content of the target document by reading and understanding the converted document.

[0017] (composition) 2 is a diagram showing an example of the configuration of the assistance system in the present disclosure. The assistance system 1 includes an assistance device 10 and a user terminal 70. The assistance device 10 is communicably connected to a storage device 60, the user terminal 70, and a generation server 80. Below, the storage device 60, the user terminal 70, the generation server 80, and the assistance device 10 will be described individually.

[0018] [Storage device] The storage device 60 stores a database 61 in which support recipient information is stored. The support recipient information is information that affects the support recipient's reading comprehension ability. The storage device 60 stores the database 61. For example, the database 61 is created and managed by the user terminal 70. The database 61 may be stored in the user terminal 70, the generation server 80, or the support device 10.

[0019] Database 61 is a database that associates identification information of support recipients with support recipient information. The identification information of support recipients is information for uniquely identifying the support recipients. For example, the identification information of support recipients includes the support recipients' national identification number (personal identification number, social security number, personal identification number, citizen ID number), name, date of birth, gender, address, telephone number, email address, etc. The support recipient information is information that represents the characteristics of the support recipients. For example, the support recipient information is the measurement values ​​of psychological tests of the support recipients. Psychological tests include intelligence tests, developmental tests, personality assessments, sociability tests, etc. For example, the support recipient information may be evaluation values ​​that represent the support recipients' lifestyle attitudes. For example, the support recipient information may be test results of brain responses or physiological responses that are a set of stimuli given to the support recipient and their response to those stimuli. For example, the support recipient information may be career information such as the support recipient's educational background and work history. The support recipient information may also include multiple evaluation items. The evaluation items will be described later.

[0020] [User terminal] The user terminal 70 is an information processing device used by a user of the support device 10. The user may be the support recipient, or a related person such as the support recipient's guardian or supervisor. The user terminal 70 has an input unit (not shown) and an output unit (not shown). The input unit is realized by an input / output interface connected to an input device such as a mouse or keyboard. The output unit is realized by an input / output interface connected to a display device such as a display. The user terminal 70 inputs items requested to be input by the support device 10 via the input unit. The user terminal 70 also displays information output from the support device 10 via the output unit.

[0021] [Generation Server] The generation server 80 stores a generation model 81. The generation model 81 is a trained model that, when a prompt including constraints and a target document are input, outputs a converted document that converts the target document to make it easier to read. The generation model 81 may be stored in the storage device 60, the user terminal 70, or the assistance device 10. Alternatively, the generation model 81 may be a general-purpose model implemented in a cloud or server accessible from the assistance device 10.

[0022] For example, the generative model 81 is a trained model constructed by machine learning using a huge number of datasets. For example, the generative model 81 is realized by a large-scale language model such as BERT (Bidirectional Encoder Representations from Transformers) or GPT (Generative Pre-trained Transformer). For example, the generative model 81 is realized by an image generation model such as VAE (variational autoencoder).

[0023] The generative model 81 may be a trained model that has been created independently. In this case, the generative model 81 is trained using a document, a prompt for converting the document, and a converted document obtained by converting the document in accordance with the prompt as training data.

[0024] [Support device] The support device 10 is a device that converts a target document into a converted document that is easier for the support recipient to read and understand, and outputs the converted converted document. For example, the support device 10 is realized by an information processing device such as a server device. The support device 10 is not limited to a server device as long as it is an information terminal that can communicate with each of the storage device 60, the user terminal 70, and the generation server 80. For example, the support device 10 may be a stationary terminal device, a notebook terminal device, a portable terminal device such as a tablet, or the like.

[0025] The support device 10 includes a support recipient information acquisition unit 11, a support evaluation value calculation unit 12, a target document acquisition unit 13, a conversion instruction reception unit 14, a prompt generation unit 15, a converted document generation unit 16, an output unit 17, and a storage unit 18.

[0026] The support recipient information acquisition unit 11 acquires support recipient information that indicates the characteristics of the support recipient from the database 61. For example, the support recipient information acquisition unit 11 acquires identification information of the support recipient from the user terminal 70. The support recipient information acquisition unit 11 searches the database 61 using the acquired identification information of the support recipient. The support recipient information acquisition unit 11 acquires support recipient information associated with the search result.

[0027] The support evaluation value calculation unit 12 calculates the support evaluation value of the support recipient using the acquired support recipient information and a calculation table for calculating the support evaluation value. The calculation table is a table for converting various information included in the support recipient information into a support evaluation value.

[0028] FIG. 3 is a diagram illustrating an example of a calculation sheet in the present disclosure. The support recipient information included in the calculation sheet may be categorized by assessment item of the test subject's abilities in an intelligence test or by assessment item of lifestyle attitudes. When the support recipient information included in the calculation sheet is categorized by assessment item, the support evaluation value calculation unit 12 calculates the total support evaluation score for each item. In the calculation sheet 181 shown in FIG. 3, the support recipient information is categorized into three assessment items: "verbal comprehension," "non-verbal comprehension," and "concentration," which are assessment items in an intelligence test. Here, verbal comprehension is an index value of the support recipient's ability to understand verbal information. verbal comprehension may be related to the support recipient's ability to understand text. non-verbal comprehension is an index value of the support recipient's ability to understand non-verbal information. non-verbal comprehension may be related to the support recipient's ability to understand non-verbal information such as diagrams. concentration is an index value of the support recipient's ability to concentrate on tasks. concentration may be related to the length of a sentence that is easy for the support recipient to read and understand. The calculation sheet does not need to be categorized by assessment item. In this case, the support evaluation value calculation unit 12 calculates the total score of the support evaluation values.

[0029] A specific example in which the support evaluation value calculation unit 12 calculates a support evaluation value will be described with reference to FIG. 3. Assume that the support recipient information acquisition unit 11 acquires an IQ of 75 as the intelligence result for the evaluation item "language comprehension" and acquires an evaluation value representing the lifestyle attitude of language comprehension that is less than the average. In this case, the support evaluation value calculation unit 12 references the evaluation item "language comprehension" in the calculation table 131 and determines that the evaluation value corresponding to the IQ of 75 is 20. Similarly, the support evaluation value calculation unit 12 determines that the evaluation value corresponding to a lifestyle attitude evaluation value that is less than the average is 15. The support evaluation value calculation unit 12 calculates 35, which is the sum of 20 and 15, as the support evaluation value for the evaluation item "language comprehension." The support evaluation value calculation unit 12 calculates support evaluation values ​​for the other evaluation items, "non-verbal comprehension" and "concentration," in the same manner.

[0030] The target document acquisition unit 13 acquires a target document that is a target for reading assistance from the user terminal 70. For example, the target document acquisition unit 13 acquires text information of the target document that is a target for reading assistance. The target document may be written in a single language or may be written in multiple languages. For example, the target document may include pictures.

[0031] The conversion instruction receiving unit 14 receives instructions regarding the conversion of the target document from the user terminal 70. The conversion instruction receiving unit 14 displays a GUI on the output unit of the user terminal 70, requesting input of instructions regarding the conversion of the target document. The conversion instruction receiving unit 14 acquires the input items as conversion instructions.

[0032] 4 is a diagram showing an example of a conversion instruction reception screen in the present disclosure. The conversion instruction reception screen W1 shown in Fig. 4 is an example of a screen for receiving instructions regarding parts of the target document that should not be changed and instructions regarding matters to be conveyed to the support recipient.

[0033] The display in data area D1 is a display for receiving instructions regarding parts of the target document that should not be modified. Data area D1 displays text information T of the target document. A user of user terminal 70 selects a sentence or passage that the user wishes not to modify from the text information T displayed in data area D1. Conversion instruction receiving unit 14 changes the display in data area D1 so that the sentence or passage selected by the user can be seen. For example, conversion instruction receiving unit 14 displays a marker M superimposed on the selected sentence or passage.

[0034] The display of data area D2 is a display for receiving instructions regarding what you want to communicate to the support recipient. Data area D2 displays a text box B1 for inputting what you want to communicate to the support recipient. The user of user terminal 70 inputs the message they want to convey to the support recipient through the target document into text box B1.

[0035] When the user terminal 70 detects that the input of the conversion instruction has been completed, it transmits the input information to the support device 10. For example, the user terminal 70 detects that the user's input has been completed by pressing button B1 on the conversion instruction reception screen W1. The conversion instruction reception unit 14 acquires the information received from the user terminal 70 as a conversion instruction.

[0036] The prompt generation unit 15 generates a prompt including a constraint condition determined in advance according to the support evaluation value calculated by the support evaluation value calculation unit 12, a conversion instruction received by the conversion instruction receiving unit 14, and a target document.

[0037] The prompt generating unit 15 identifies a predetermined constraint condition according to the support evaluation value using a constraint condition determination table. The constraint condition determination table is a table that associates the support evaluation value with the constraint condition corresponding to that value.

[0038] FIG. 5 is a diagram illustrating an example of a constraint condition determination table in the present disclosure. In the constraint condition determination table 182, a constraint condition is determined for each evaluation item, but this is just one example. For example, if the support evaluation value calculation unit 12 does not calculate a support evaluation value for each evaluation item, the constraint condition determination table does not include information about the evaluation item. The prompt generation unit 15 identifies constraint conditions corresponding to the support evaluation value calculated by the support evaluation value calculation unit 12. For example, if the support evaluation value calculated by the support evaluation value calculation unit 12 is 5, the prompt generation unit 15 identifies the following instructions as constraint conditions: "output in text format, annotate technical terms, highlight parts related to 'matters to be conveyed to the reader,' and do not modify the original document."

[0039] The constraints may include a document format specification based on the support evaluation value. As shown in FIG. 5, the constraints may include the level of kanji to be used in the converted document. For example, the kanji level is set according to a hierarchy such as elementary school completion level, junior high school completion level, or high school completion level. For example, the kanji level is set according to the level of a kanji proficiency test. As shown in FIG. 5, the constraints may also include the number of characters in a sentence constituting the converted document. As shown in FIG. 5, the constraints may also include the media type of the converted document. The media type indicates a classification of output media such as "text," "audio," or "picture book." Text indicates output information composed of text information. Audio indicates output information composed of audio information. Picture book indicates output information in which relevant illustrations are inserted in addition to text. The media type is not limited to the above and may also be manga or video. As shown in FIG. 5, the constraints may also include an instruction to clearly indicate the parts that have been modified. As shown in FIG. 5, the constraints may also include an instruction to output the original document, the target document, along with the converted document. The constraints may also include specifications regarding style, such as line spacing. In addition, the constraints may include other information such as character size, font, etc.

[0040] For example, assume that the support recipient information acquired by the support recipient information acquisition unit 11 includes the evaluation items of "language comprehension" and "non-language comprehension." In this case, the prompt generation unit 15 may generate a prompt including a medium type of the converted document as a constraint, which is predetermined based on a result of comparing the support evaluation value of language comprehension with the support evaluation value of non-language comprehension. For example, if the support evaluation value of the support recipient's language comprehension is higher than the support evaluation value of non-language comprehension, the constraint condition determination table 182 may store a constraint condition specifying that the type of output medium is text. Furthermore, if the support evaluation value of the support recipient's language comprehension is lower than the support evaluation value of non-language comprehension, the constraint condition determination table 182 may store a constraint condition specifying that the type of output medium is a picture book. In this case, the prompt generation unit 15 compares the support evaluation value of language comprehension with the support evaluation value of non-language comprehension and determines a constraint condition based on the comparison result.

[0041] When the support recipient information acquired by the support recipient information acquisition unit 11 includes the evaluation item of "concentration," the prompt generation unit 15 generates a prompt including a predetermined constraint according to the support evaluation value of concentration. In this case, the constraint is the number of characters in one sentence that makes up the converted document. Alternatively, the constraint may be a specification regarding style, such as line spacing.

[0042] The conversion instruction receiving unit 14 may receive an instruction to change the constraint condition identified by the prompt generating unit 15. For example, the instruction to change the constraint condition is specified via a conversion instruction receiving screen.

[0043] FIG. 6 illustrates an example of a conversion instruction accepting screen for accepting an instruction to change a constraint. The conversion instruction accepting unit 14 displays the conversion instruction accepting screen W2 on the display unit of the user terminal 70. The conversion instruction accepting screen W2 displays a list of constraints identified by the prompt generating unit 15. The conversion instruction accepting screen W2 also includes options other than the constraints identified by the prompt generating unit. For example, in the example of FIG. 6, the conversion instruction accepting screen displays "picture book," which is identified as the medium type, as well as other options, "text" and "audio." For example, if the GUI is a radio button, the button corresponding to the constraint identified by the prompt generating unit 15 is initially selected and displayed. For example, if the GUI is a drop-down list, the constraint identified by the prompt generating unit 15 is initially displayed. This allows the user to understand the constraints identified by the prompt generating unit 15 even when other options are displayed. The user can change the constraints identified by the prompt generating unit 15 as needed via the input unit of the user terminal 70. When the user terminal 70 detects that the input of the conversion instruction has been completed, it transmits the input information to the assistance device 10. For example, the user terminal 70 detects that the user's input has been completed when button B3 is pressed on the conversion instruction reception screen W2. The conversion instruction reception unit 14 acquires the information received from the user terminal 70. The prompt generation unit changes the identified constraint condition using the information received by the conversion instruction reception unit.

[0044] The prompt generator 15 generates a prompt including a constraint condition, and the prompt includes an instruction for outputting desired output information from the generative model.

[0045] Fig. 7 is a diagram showing an example of a prompt generated by the prompt generation unit 15 in the present disclosure. The prompt generated by the prompt generation unit 15 will be described with reference to Fig. 7. The prompt generated by the prompt generation unit 15 includes a command, a constraint, and an input document.

[0046] The instruction is a statement that specifies the role of the generative model 81 and the type of output data that the generative model 81 should generate. In the example of Figure 7, the instruction states, "You are an explainer. Please convert the input document so that it is easy for the reader to understand, in accordance with the following constraints." Of this statement, the statement "You are an explainer" corresponds to the specification of the role of the generative model 81. Also, of this statement, the statement "Please convert the input document so that it is easy for the reader to understand, in accordance with the following constraints" corresponds to the specification of the type of output data that the generative model 81 should generate. The items to be written in the instruction are predetermined.

[0047] The constraint conditions are command statements stating instructions for controlling output data. The constraint conditions include constraint conditions determined in advance according to the support evaluation value calculated by the support evaluation value calculation unit 12 and constraint conditions including conversion instructions. In the example shown in FIG. 7, the constraint conditions include data D3 and data D4. The items described in data D3 correspond to the constraint conditions determined in advance according to the support evaluation value calculated by the support evaluation value calculation unit 12. The items described in data D4 correspond to constraint conditions including conversion instructions.

[0048] The input document is text information of the target document, and includes the text information of the target document acquired by the target document acquisition unit 13.

[0049] The transferred document generation unit 16 generates a transferred document using the generative model 81. The transferred document generation unit 16 inputs the prompt generated by the prompt generation unit 15 into the generative model 81. In response to the input prompt, the generative model 81 generates a transferred document. The transferred document is a target document that has been converted to content that is easy for the support recipient to read and understand. The transferred document generation unit 16 outputs the transferred document generated by the generative model 81.

[0050] The output unit 17 outputs the transferred document generated by the transferred document generation unit 16 to the user terminal 70. The transferred document output by the output unit 17 is displayed on the output unit of the user terminal 70.

[0051] The storage unit 18 stores various types of information used in the processing of the support device 10. Specifically, the storage unit 18 stores a calculation table and a constraint condition determination table. The storage unit 18 may store information other than the above-mentioned information, such as information temporarily used in the processing of the support device 10.

[0052] (operation) Next, the operation of the support device 10 according to this embodiment will be described with reference to Fig. 8. Fig. 8 is a flowchart for explaining the operation of the support device according to the present disclosure. For example, the support device 10 starts processing by receiving identification information of the support recipient from the user terminal 70.

[0053] First, the support recipient information acquisition unit 11 acquires support recipient information from the database 61 (step S11). Identification information of the support recipient and support recipient information are associated in the database 61. The support recipient information acquisition unit 11 searches the database 61 using the identification information of the support recipient, and acquires support recipient information associated with the search result.

[0054] Next, the support evaluation value calculation unit 12 calculates the support evaluation value of the support recipient using the support recipient information acquired in step S11 and a calculation table for calculating the support evaluation value (step S12). The calculation table is a table for converting various information included in the support recipient information into a support evaluation value.

[0055] Next, the target document acquisition unit 13 acquires text information of the target document that is the target of reading support from the user terminal 70 (step S13).

[0056] Next, the conversion instruction receiving unit 14 receives instructions regarding the conversion of the target document from the user terminal 70 (step S14). For example, the conversion instruction receiving unit 14 receives instructions regarding parts of the target document that should not be changed. For example, the conversion instruction receiving unit 14 receives instructions regarding matters that should be communicated to the support recipient.

[0057] Next, the prompt generation unit 15 identifies a predetermined constraint condition according to the support evaluation value calculated in step S12 (step S15). The prompt generation unit identifies a predetermined constraint condition according to the support evaluation value by using a constraint condition determination table stored in the storage unit 18. The conversion instruction receiving unit 14 may receive an instruction to change the constraint condition identified by the prompt generation unit 15.

[0058] Next, the prompt generation unit 15 generates a prompt including the constraints identified in step S15 and the conversion instructions received in step S14 (step S16). The prompt generated by the prompt generation unit 15 includes an instruction, constraints, and an input document. The instruction is a statement specifying the role of the generative model 81 and the type of output data that the generative model 81 should generate. The constraints are a statement specifying instructions for controlling the output data. The input document is text information of the target document.

[0059] Next, the transferred document generation unit 16 inputs the prompt generated in step S16 into the generative model 81 to generate a transferred document (step S17). The transferred document is a target document that has been converted in accordance with the prompt.

[0060] Next, the output unit 17 outputs the transferred document generated by the transferred document generation unit 16 to the user terminal 70 (step S18). After completing the process of step S18, the support device 10 ends the series of processes described above.

[0061] As described above, the assistance device of this embodiment includes an assistance recipient information acquisition unit, an assistance evaluation value calculation unit, a target document acquisition unit, a prompt generation unit, a converted document generation unit, and an output unit. The assistance recipient information acquisition unit acquires assistance recipient information that indicates the characteristics of the assistance recipient who is the target of reading assistance. The assistance evaluation value calculation unit calculates an assistance evaluation value that indicates the degree of reading assistance for the assistance recipient according to the assistance recipient information. The target document acquisition unit acquires a target document that is the target of reading assistance. The prompt generation unit generates a prompt that includes a constraint according to the assistance evaluation value and the target document. The converted document generation unit generates output information that is output from the trained generative model in response to input of the prompt as a converted document. The output unit outputs the generated converted document.

[0062] By having the above-described configuration, the assistance device in this embodiment can provide reading assistance according to the characteristics of the assistance recipient. The assistance device in this embodiment acquires assistance recipient information, which is information representing the characteristics of the assistance recipient, calculates an assistance evaluation value based on the assistance recipient information, generates a prompt including predetermined constraints according to the assistance evaluation value, and inputs the prompt into a generative model to generate a converted document. In other words, the assistance device in this embodiment can provide reading assistance according to the characteristics of the assistance recipient. In other words, the assistance device in this embodiment can provide reading assistance according to the characteristics of the assistance recipient.

[0063] In the assistance device according to one aspect of this embodiment, the constraint conditions include the specification of a document format according to the assistance evaluation value. For example, the constraint conditions include the level of kanji to be used in the converted document. The constraint conditions are determined in advance according to the assistance evaluation value. One of the factors that hinders the reader (support recipient) from comprehending the document is the inclusion of difficult kanji in the target document. According to the assistance device according to one aspect of this embodiment, the level of kanji included in the converted document, which is output information, can be changed according to the characteristics of the support recipient. In other words, according to the assistance device according to this embodiment, reading support can be provided according to the characteristics of the support recipient.

[0064] In the assistance device according to one aspect of this embodiment, the constraint conditions include the specification of a document format according to the assistance evaluation value. For example, the constraint conditions include the number of characters in one sentence that constitutes the converted document. The constraint conditions are determined in advance according to the assistance evaluation value. One of the factors that hinders reading comprehension by a reader (person receiving assistance) is the length of one sentence that constitutes the target document. According to the assistance device according to one aspect of this embodiment, the number of characters in one sentence that constitutes the converted document, which is output information, can be changed according to the characteristics of the person receiving assistance. In other words, the assistance device according to this embodiment can provide reading comprehension support according to the characteristics of the person receiving assistance.

[0065] In the support device according to one aspect of this embodiment, the support recipient information includes at least one of a psychological test measurement value of the support recipient or an evaluation value representing the support recipient's lifestyle. The psychological test and lifestyle include a plurality of evaluation items. The support evaluation value calculation unit calculates a support evaluation value for each evaluation item. The prompt generation unit generates a prompt including a predetermined constraint condition and a target document according to the support evaluation value of the evaluation item. The evaluation item is an item that may hinder the reader's reading comprehension. The support device according to one aspect of this embodiment changes the prompt according to the support evaluation value of each evaluation item of the support recipient. In other words, the support device according to one aspect of this embodiment can provide reading comprehension support according to the characteristics of the support recipient.

[0066] In an assistance device according to one aspect of the present embodiment, evaluation items include language comprehension, which is the ability to understand language information, and non-verbal comprehension, which is the ability to understand non-verbal information. The prompt generation unit generates a prompt that includes a medium type of the converted document as a constraint, which is predetermined based on a result of comparing an assistance evaluation value for language comprehension with an assistance evaluation value for non-verbal comprehension. The medium types include text, audio, and picture books. The converted document generation unit generates a converted document of the medium type included in the prompt. Some assistance recipients find it easier to understand content when explained using language information such as text, while others find it easier to understand content when explained using non-verbal information such as drawings. The assistance recipient's language comprehension and non-verbal comprehension are one indicator of whether the assistance recipient can understand language information or non-verbal information more easily. The assistance device according to one aspect of the present embodiment generates a prompt that includes a medium type of the converted document as a constraint, which is predetermined based on a result of comparing an assistance evaluation value for language comprehension with an assistance evaluation value for non-verbal comprehension. The medium types include text, audio, and picture books. In other words, the support device of this embodiment changes the medium type of the output information depending on the reader's linguistic comprehension ability and non-linguistic comprehension ability. In other words, the support device of this embodiment can provide reading support that is tailored to the characteristics of the support recipient.

[0067] The support device of one aspect of this embodiment further includes a conversion instruction receiving unit that receives instructions regarding the conversion of the target document. The instructions are instructions regarding parts of the target document that should not be modified. The prompt generation unit generates a prompt that further includes the instructions. Modifying the content of the document may change the meaning of the original document. By receiving instructions regarding parts of the target document that should not be modified and including the instructions in the prompt, the support device of one aspect of this embodiment can convert the target document without modifying the parts that are instructed.

[0068] In the assistance device according to one aspect of this embodiment, the instructions received by the conversion instruction receiving unit further include instructions regarding matters to be conveyed to the assistance recipient. The prompt generating unit generates a prompt that further includes the instructions. Depending on the target document, the matters to be conveyed to the reader may not be explicitly stated. By receiving instructions regarding the matters to be conveyed in the target document and including the instructions in the prompt, the assistance device according to one aspect of this embodiment can convert the target document to include the matters to be conveyed. This makes it possible to provide reading comprehension support to the reader.

[0069] (Application example) The effects of applying the support system according to this embodiment to specific scenes will be described below. In this application example, three application scenes will be described, but these are just examples. The support system of this embodiment can also be applied to reading comprehension support in other scenes, such as job hunting support and rehabilitation support.

[0070] (Application example 1) In this application example, the effect when the support recipient is a person with borderline intelligence is described. Borderline intelligence indicates that the measured IQ is on the border between the range judged to be normal and the range judged to be intellectually disabled.

[0071] The support recipient information includes at least one of the measured values ​​of the psychological test of the support recipient or the evaluation value representing the support recipient's lifestyle. Therefore, the support device of this embodiment can provide reading support to people with borderline intelligence.

[0072] Even if a person with borderline intelligence can partially understand a target document, they may not be able to grasp the overall meaning of the document or the content the writer is trying to convey through the document. The conversion instruction receiving unit of this embodiment receives instructions regarding the matters to be conveyed to the support recipient. The prompt generating unit generates a prompt including the matters to be conveyed to the support recipient. The converted document generating unit converts the target document so that it includes the matters to be conveyed to the support recipient. This makes it possible to provide the reader with a converted document that includes the content to be conveyed to the reader, even if the matters to be conveyed to the reader are not clearly stated in the target document. In other words, the support device of this application example can provide reading comprehension support that is tailored to the characteristics of a person with borderline intelligence.

[0073] Some individuals with borderline intelligence may not have all of their intelligence at the borderline level. For example, among the evaluation items tested in an intelligence test, only language comprehension may be at the borderline level, while the other evaluation items may be in the normal range. In this case, if the target document is converted into non-verbal information, the support recipient can easily understand the content. The support device of this embodiment generates a prompt that includes the media type of the converted document as a constraint, which is predetermined based on the result of comparing the support evaluation value for language comprehension and the support evaluation value for non-verbal comprehension. In other words, the support device of this embodiment can generate output information of a media type that is easy for the reader to understand. In other words, the support device of this application example can provide reading support tailored to the characteristics of individuals with borderline intelligence.

[0074] (Application example 2) In this application example, the effect when the target document is a technical book will be described. Technical books are written assuming that the readers are experts in the field, so they are difficult for people who are not experts in the field to read and understand. However, even people who are not experts in the field may need to read technical books. In such cases, the effect when the support system of this embodiment is applied will be described.

[0075] The prompts input to the generative model are determined by a support evaluation value based on whether the support recipient has a learning history in the field. For example, a constraint condition of "include technical terminology annotations" is associated with a person who has no learning history in the field. This allows even a person who is well-versed in the specialized field to understand the specialized book. In other words, the support device of this application example can provide reading support tailored to the characteristics of a person who does not have knowledge in the field of the specialized book.

[0076] Since technical books preferably contain detailed descriptions of specialized matters, it is desirable to avoid alterations as much as possible. The support device of this embodiment accepts an instruction to specify the portions of the target document that should not be altered. The portions that receive this instruction are not altered. This allows the target document to be converted to be easier to read without altering the portions that should not be altered. In other words, the support device of this embodiment can provide reading support to those who do not have knowledge in the field of the technical book.

[0077] (Application example 3) In this application example, the effect when the support recipient is a child will be described. Children often have a certain level of kanji that they can understand depending on their age. The support device of this embodiment acquires the age of the support recipient as support recipient information, calculates a support evaluation value depending on the age, and generates a prompt including the level of kanji to be used in the converted document, which is predetermined depending on the support evaluation value. As a result, the support device of this embodiment can provide reading support that is suited to the characteristics of the support recipient, even when the support recipient is a child.

[0078] (Second embodiment) In this embodiment, a support device 20 will be described, which has a simplified configuration of the support device according to the first embodiment. In the following description, the same parts as those in the first embodiment will not be described as appropriate.

[0079] (composition) Next, the configuration of the assistance device according to this embodiment will be described with reference to the drawings. Fig. 9 is a block diagram showing an example of the configuration of the assistance device according to the present disclosure. The assistance device 20 includes an assistance recipient information acquisition unit 21, an assistance evaluation value calculation unit 22, a target document acquisition unit 23, a conversion instruction acceptance unit 24, a prompt generation unit 25, a converted document generation unit 26, and an output unit 27.

[0080] The support recipient information acquisition unit 21 acquires support recipient information that indicates the characteristics of the support recipient who is the target of reading support. The support evaluation value calculation unit 22 calculates a support evaluation value that indicates the degree of reading support for the support recipient according to the support recipient information. The target document acquisition unit 23 acquires a target document that is the target of reading support. The prompt generation unit 25 generates a prompt that includes a constraint condition according to the support evaluation value and the target document. The converted document generation unit 26 generates output information output from the trained generative model in response to the input of the prompt as a converted document. The output unit 27 outputs the generated converted document.

[0081] (operation) Next, an example of the operation of the assistance device according to this embodiment will be described with reference to Fig. 10. Fig. 10 is a diagram showing an example of the operation of the assistance device in the present disclosure.

[0082] First, the support recipient information acquisition unit 21 acquires support recipient information that indicates the characteristics of the support recipient who is the target of reading support (step S21).

[0083] Next, the support evaluation value calculation unit 22 calculates a support evaluation value indicating the degree of reading support for the support recipient, according to the support recipient information (step S22).

[0084] Next, the target document acquisition unit 23 acquires a target document that is the target of reading support (step S23).

[0085] Next, the prompt generating unit 25 generates a prompt including a constraint condition according to the support evaluation value and a target document (step S24).

[0086] Next, the transferred document generation unit 26 generates output information output from the trained generative model in response to the input of the prompt as a transferred document (step S25).

[0087] Next, the output unit 27 outputs the generated transferred document (step S26).

[0088] The assistance device of this embodiment calculates an assistance evaluation value according to the characteristics of the assistance recipient, generates a prompt including a predetermined constraint and a target document according to the calculated assistance evaluation value, and generates a converted document by inputting the prompt into a trained model. In other words, the assistance device of this embodiment can provide reading comprehension assistance according to the characteristics of the assistance recipient.

[0089] (Hardware configuration) The functions of each of the components according to the embodiments of the present disclosure described above can be realized not only as hardware but also as a computer device or firmware under program control.

[0090] 11 is a diagram showing an example of a hardware configuration in which the assistance device according to the present disclosure is realized by a computer device 90 including a processor. The assistance device according to each embodiment is realized by the computer device 90. As shown in FIG. 11, the computer device 90 includes a processor 91, a memory 92, a storage device 93 such as a hard disk for storing programs, an input / output interface 94 for connecting input devices and output devices, and a communication interface 95 for connecting to a network.

[0091] The processor 91 loads a program (instructions) stored in a storage device 93 or the like into the memory 92. For example, the program is a software program for executing the control and processing in the present disclosure. The processor 91 executes the program loaded into the memory 92. The processor 91 executes the program to execute the control and processing in the present disclosure.

[0092] The storage device 93 may be, for example, an optical disk, a flexible disk, a magneto-optical disk, an external hard disk, or a semiconductor memory. Some storage media in the storage device are non-volatile storage devices, and the programs are recorded therein. The programs may also be downloaded from an external computer (not shown) connected to a communication network.

[0093] The input device connected to the input / output interface 94 is realized by, for example, a mouse or a keyboard, and is used for input operations. Similarly, the output device connected to the input / output interface 94 is realized by, for example, a display, and is used for displaying and checking output results.

[0094] Although the present disclosure has been described above with reference to the embodiments, the present disclosure is not limited to the above-described embodiments. Various modifications that can be understood by those skilled in the art can be made to the configuration and details of the present disclosure within the scope of the present disclosure. Furthermore, each embodiment can be combined with other embodiments as appropriate.

[0095] Some or all of the above embodiments can also be described as follows:

[0096] (Appendix 1) a support recipient information acquisition means for acquiring support recipient information representing characteristics of a support recipient who is a target of reading support; a support evaluation value calculation means for calculating a support evaluation value indicating a degree of reading support for the support recipient according to the support recipient information; a target document acquisition means for acquiring a target document that is a target of reading support; a prompt generating means for generating a prompt including a constraint condition according to the support evaluation value and the target document; a transfer document generation means for generating output information output from the trained generative model in response to the input of the prompt as a transfer document; and an output means for outputting the generated transferred document.

[0097] (Appendix 2) The support evaluation value calculation means 2. The support device according to claim 1, wherein the support evaluation value of the support recipient is calculated using a calculation table in which the support recipient information and the support evaluation value are associated with each other.

[0098] (Appendix 3) 3. The assistance device according to claim 1, wherein the constraint condition includes a specification of a document format according to the assistance evaluation value.

[0099] (Appendix 4) The support recipient information includes at least one of a measured value of a psychological test of the support recipient or an evaluation value representing the support recipient's lifestyle, The psychological test and the lifestyle attitude include a plurality of evaluation items, the support evaluation value calculation means calculates the support evaluation value for each of the evaluation items; 4. The assistance device according to claim 1, wherein the prompt generation means generates the prompt including a predetermined constraint condition according to the assistance evaluation value of the evaluation item and the target document.

[0100] (Appendix 5) The evaluation items include language comprehension, which is the ability to understand linguistic information, and non-verbal comprehension, which is the ability to understand non-verbal information, the prompt generation means generates the prompt, which is predetermined in accordance with a result of comparing the support evaluation value of the language comprehension ability with the support evaluation value of the non-language comprehension ability, and which includes a medium type of the converted document as a constraint; the media types include text, audio, and picture books; 5. The support device according to claim 4, wherein the transfer document generation means generates the transfer document of the medium type included in the prompt.

[0101] (Appendix 6) further comprising a conversion instruction receiving means for receiving an instruction regarding the conversion of the target document; The instruction is an instruction regarding a portion of the target document that should not be altered, 6. The assistance device according to claim 1, wherein the prompt generating means generates the prompt further including the instruction.

[0102] (Appendix 7) 7. The support device according to claim 6, wherein the instructions further include instructions about matters to be conveyed to the support recipient.

[0103] (Appendix 8) An assistance device according to any one of Supplementary Notes 1 to 7; a user terminal having an output unit for displaying the converted document; A support system comprising:

[0104] (Appendix 9) The computer Acquire support recipient information that indicates the characteristics of the support recipient who is the target of reading support; calculating a support evaluation value indicating a degree of reading support for the support recipient according to the support recipient information; Obtain the target document that is the subject of reading support; generating a prompt including a constraint condition according to the support evaluation value and the target document; Generate output information output from the trained generative model in response to the input of the prompt as a converted document; A support method for outputting the generated transfer document.

[0105] (Appendix 10) On the computer, A process of acquiring support recipient information that indicates characteristics of a support recipient who is a target of reading support; A process of calculating a support evaluation value indicating the degree of reading support for the support recipient according to the support recipient information; A process of acquiring a target document that is a target of reading assistance; generating a prompt including a constraint condition according to the support evaluation value and the target document; A process of generating output information output from the trained generative model in response to input of the prompt as a converted document; and a program for executing a process of outputting the generated converted document.

[0106] In addition, in the above appendices, some or all of the configurations described in appendices 2 to 7 that are dependent on appendices 1 may also be dependent on appendices 9 and 10 in the same dependent relationship as appendices 2 to 7. Furthermore, not limited to appendices 1, 9, and 10, some or all of the configurations described as appendices may be made dependent on various hardware, software, various recording means for recording software, or systems, within the scope of each of the above-mentioned embodiments. [Explanation of symbols]

[0107] 10, 20 Support equipment 11, 21 Support recipient information acquisition department 12, 22 Support evaluation value calculation unit 13, 23 Target document acquisition section 14 Conversion instruction reception unit 15, 25 Prompt generation section 16, 26 Transformation document generation unit 17, 27 Output section 18 Memory section 60 Storage device 61 databases 70 User terminal 80 Generation Server 81 Generative Model 90 Computer Equipment 91 processors 92 memory 93 Storage device 94 Input / Output Interface 95 Communication Interface

Claims

1. a support recipient information acquisition means for acquiring support recipient information representing characteristics of a support recipient who is a target of reading support; a support evaluation value calculation means for calculating a support evaluation value indicating a degree of reading support for the support recipient according to the support recipient information; a target document acquisition means for acquiring a target document that is a target of reading support; a prompt generating means for generating a prompt including a constraint condition according to the support evaluation value and the target document; a transfer document generation means for generating output information output from the trained generative model in response to the input of the prompt as a transfer document; and an output means for outputting the generated transferred document.

2. The support evaluation value calculation means The support device according to claim 1 , wherein the support evaluation value of the support recipient is calculated using a calculation table in which the support recipient information and the support evaluation value are associated with each other.

3. The assistance device according to claim 1 , wherein the constraint condition includes a specification of a document format according to the assistance evaluation value.

4. The support recipient information includes at least one of a measured value of a psychological test of the support recipient or an evaluation value representing the support recipient's lifestyle, The psychological test and the lifestyle attitude include a plurality of evaluation items, the support evaluation value calculation means calculates the support evaluation value for each of the evaluation items; 2. The assistance device according to claim 1, wherein the prompt generating means generates the prompt including a constraint condition determined in advance according to the assistance evaluation value of the evaluation item and the target document.

5. The evaluation items include language comprehension, which is the ability to understand linguistic information, and non-verbal comprehension, which is the ability to understand non-verbal information, the prompt generation means generates the prompt, which is predetermined in accordance with a result of comparing the support evaluation value of the language comprehension ability with the support evaluation value of the non-language comprehension ability, and which includes a medium type of the converted document as a constraint; the media types include text, audio, and picture books; 5. The support device according to claim 4, wherein the transfer document generating means generates the transfer document for the medium type included in the prompt.

6. further comprising a conversion instruction receiving means for receiving an instruction regarding the conversion of the target document; The instruction is an instruction regarding a portion of the target document that should not be altered, The assistance device according to claim 1 , wherein the prompt generating means generates the prompt further including the instruction.

7. The support device according to claim 6 , wherein the instructions further include instructions about matters to be conveyed to the support recipient.

8. An assistance device according to any one of claims 1 to 7; a user terminal having an output unit for displaying the converted document; A support system comprising:

9. The computer Acquire support recipient information that indicates the characteristics of the support recipient who is the target of reading support; calculating a support evaluation value indicating a degree of reading support for the support recipient according to the support recipient information; Obtain the target document that is the subject of reading support; generating a prompt including a constraint condition according to the support evaluation value and the target document; Generate output information output from the trained generative model in response to the input of the prompt as a converted document; A support method for outputting the generated transfer document.

10. On the computer, A process of acquiring support recipient information that indicates characteristics of a support recipient who is a target of reading support; A process of calculating a support evaluation value indicating the degree of reading support for the support recipient according to the support recipient information; A process of acquiring a target document that is a target of reading assistance; generating a prompt including a constraint condition according to the support evaluation value and the target document; A process of generating output information output from the trained generative model in response to input of the prompt as a converted document; and a program for executing a process of outputting the generated converted document.

Citation Information

Patent Citations

  • Assist system, assist method, and assist program

    WO2022070747A1