Learning assistance device, program, method, and system

Abstract

The present invention relates to a learning assistance device, a program, a method, and a system. A learning assistance device 16 comprises: an output instruction unit 50 that instructs an output device (24) to present to a user, at a presentation timing for target content, learning presentation information (110, 120, 130) which relates to learning the target content; a data acquisition unit 40 that acquires result data (D2) which is obtained as hand-written input by the user with respect to the learning presentation information (110, 120, 130) presented using the output device (24); an retention degree calculation unit (82, 92) that performs an analysis process with respect to the result data (D2) and calculates the degree of retention of user knowledge relating to the target content; and a timing setting unit (84) that sets the next presentation timing for the target content in accordance with the calculated degree of retention.

Description

Learning support device, program, method, and system

[0001] The present invention relates to a learning support device, a program, a method, and a system.

[0002] 2. Description of the Related Art Conventionally, there have been known techniques for supporting user learning using computer systems.

[0003] Japanese Patent No. 6592657 discloses a vocabulary acquisition training system for learning related words from one word in a chain reaction, in accordance with the learner's learning level, learning purpose, and field of use of the language. Specifically, Patent Document 1 describes that comprehensive training of the four language skills of reading, listening, writing, and speaking can be carried out by practicing words in sentences that are suitable for the learner, thereby enabling the learner to acquire many words effectively and efficiently.

[0004] Patent No. 6592657

[0005] However, while the system disclosed in Patent Publication No. 6592657 allows for the continuous learning of multiple learning contents, it may not take into account memory phases based on neuroscience findings.

[0006] The present invention has been made in consideration of these problems, and its purpose is to provide a learning support device, program, method, and system that can improve the degree to which knowledge is retained through a learning format that presents multiple learning contents to users.

[0007] A learning support device in a first aspect of the present invention comprises an output instruction unit that instructs an output device to present learning presentation information related to learning of a target content, which is at least one learning content among a plurality of learning contents, to a user at the timing of presentation of the target content; a data acquisition unit that acquires product data, which is data indicating a product obtained by the user's handwritten input on the learning presentation information presented using the output device, and which includes stroke data for describing strokes; a fixation calculation unit that performs an analytical process on the product data acquired by the data acquisition unit to calculate the fixation of the user's knowledge regarding the target content; and a timing setting unit that sets the next presentation timing of the target content according to the fixation calculated by the fixation calculation unit.

[0008] In a second aspect of the present invention, the learning support program causes one or more computers to execute the following steps: an instruction step of instructing an output device to present learning presentation information related to learning of a target content, which is at least one learning content among multiple learning contents, to a user at the timing of presentation of the target content; an acquisition step of acquiring product data, which is data indicating a product obtained by the user's handwritten input on the learning presentation information presented using the output device, and which includes stroke data for describing strokes; a calculation step of performing an analysis process on the acquired product data to calculate the user's level of knowledge retention regarding the target content; and a setting step of setting the next presentation timing of the target content based on the calculated level of retention.

[0009] In a third aspect of the present invention, the learning support method is performed by one or more computers, which perform the following steps: an instruction step of instructing an output device to present learning presentation information related to learning of a target content, which is at least one learning content among a plurality of learning contents, to a user at the timing of presentation of the target content; an acquisition step of acquiring product data, which is data indicating a product obtained by the user's handwritten input on the learning presentation information presented using the output device, and which includes stroke data for describing strokes; a calculation step of performing an analysis process on the acquired product data to calculate the user's level of knowledge retention regarding the target content; and a setting step of setting the next presentation timing of the target content according to the calculated level of retention.

[0010] A learning support system in a fourth aspect of the present invention comprises a user terminal that generates product data, which is data indicating a product obtained by a user's handwritten input and includes stroke data for describing strokes, and a learning support device configured to be able to communicate with the user terminal, wherein the learning support device comprises an output instruction unit that instructs an output device to present learning presentation information related to learning of target content, which is at least one learning content among a plurality of learning contents, to the user at the presentation timing of the target content, a data acquisition unit that acquires from the user terminal the product data corresponding to the learning presentation information presented using the output device, a fixation calculation unit that performs an analytical process on the product data acquired by the data acquisition unit to calculate the fixation of the user's knowledge regarding the target content, and a timing setting unit that sets the next presentation timing of the target content according to the fixation calculated by the fixation calculation unit.

[0011] According to the present invention, the degree to which knowledge is retained can be improved through a learning format in which a plurality of learning contents are presented to the user.

[0012] 12. FIG. 13 is an overall configuration diagram of a learning assistance system according to an embodiment of the present invention. FIG. 14 is a block diagram showing an example of the configuration of the learning assistance device in FIG. 1. FIG. 15 is a detailed functional block diagram of the tag generation unit shown in FIG. 2. FIG. 16 is a detailed functional block diagram of the exercise processing unit shown in FIG. 2. FIG. 17 is a detailed functional block diagram of the evaluation processing unit shown in FIG. 2. FIG. 5 is a diagram showing an example of a network structure possessed by a learning device included in the index generation unit of FIG. 5. FIG. 6 is a sequence diagram showing the operation of updating a knowledge graph. FIG. 7 is a diagram showing an example of a collection of strokes displayed on the display unit of a learner terminal. FIG. 8 is a diagram showing an example of the data structure of digital ink. FIG. 9 is a diagram showing the result of stroke classification. FIG. 10 is a sequence diagram showing the operation of supporting a learner. FIG. 11 is a diagram showing a first example of an exercise screen displayed on the display unit of a learner terminal. FIG. 12 is a diagram showing a second example of an exercise screen displayed on the display unit of a learner terminal. FIG. 13 is a diagram showing a third example of an exercise screen displayed on the display unit of a learner terminal. FIG. 14 is a diagram showing a first example of the data structure possessed by the learning outcome information of FIG. 2. FIG. 15 is a first flowchart showing the operation of generating question data (step SP36 of FIG. 12). FIG. 16 is a diagram showing the Evins-House forgetting curve. FIG. 17 is a diagram showing an example of the process of knowledge consolidation. FIG. 18 is a diagram showing a first example of a method for setting presentation timing. 12A is a diagram showing a second example of the data structure of the learning outcome information of FIG. 2. FIG. 13B is a second flowchart relating to the operation of generating question data (step SP36A of FIG. 12). FIG. 14C is a diagram showing an example of time change in retention based on a forgetting model. FIG. 15A is a diagram showing a second example of a method for setting presentation timing.

[0013] Hereinafter, an embodiment of the present invention will be described with reference to the accompanying drawings. To facilitate understanding of the description, the same components in each drawing will be denoted by the same reference numerals as much as possible, and duplicate explanations will be omitted. Furthermore, the term "part" may be replaced with other terms such as unit, module, device, or element.

[0014] [Configuration of Learning Support System 10] <Overall Configuration> Fig. 1 is a diagram showing the overall configuration of a learning support system 10 according to one embodiment of the present invention. The learning support system 10 is provided to provide a "learning support service" for supporting learners in various types of learning, including compulsory education, self-development, or skill training. This learning may be either "input learning" aimed at inputting knowledge, or "output learning" aimed at outputting knowledge. In this embodiment, an explanation will be given using "exercises," which is one aspect of output learning, as an example.

[0015] Specifically, this learning support system 10 includes one or more learner terminals 12 (corresponding to "user terminals"), one or more electronic pens 14, and a learning support device 16. Each learner terminal 12 and learning support device 16 are connected to each other so as to be able to communicate with each other via a network NT.

[0016] The learner terminal 12 is a computer owned by a user (here, a learner) who uses the learning support service, and is composed of a tablet, smartphone, personal computer, wearable device, etc. Specifically, the learner terminal 12 is composed of a processor 21, a memory 22, a communication unit 23, a display unit 24, and a touch sensor 25.

[0017] The processor 21 is configured with an arithmetic processing device including a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), and an MPU (Micro-Processing Unit). The processor 21 reads out programs and data stored in the memory 22, and performs a generation process to generate ink data (hereinafter also referred to as digital ink) that describes content, a rendering process to display content indicated by the digital ink, and the like.

[0018] The memory 22 stores programs and data necessary for the processor 21 to control each component. The memory 22 is non-transitory and is composed of a computer-readable storage medium. Here, the computer-readable storage medium is composed of [1] a storage device such as a hard disk drive (HDD) or solid state drive (SSD) built into the computer system, or [2] a portable medium such as a magneto-optical disk, a read-only memory (ROM), a compact disc (CD)-ROM, or a flash memory.

[0019] The communication unit 23 has a communication function for performing wired or wireless communication with an external device, whereby the learner terminal 12 exchanges various data, including question data D1 and answer data D2, which will be described later with reference to FIG. 2, with the learning support device 16, for example.

[0020] The display unit 24 can visibly display content including images or videos, and is configured, for example, by a liquid crystal panel, an organic EL (Electro-Luminescence) panel, or electronic paper. By making the display unit 24 flexible, the learner can perform various writing operations while keeping the touch surface of the learner terminal 12 curved or bent.

[0021] The touch sensor 25 is, for example, a capacitance-type planar sensor having a plurality of sensor electrodes arranged in a plane. The touch sensor 25 includes, for example, a plurality of X-line electrodes for detecting the position on the X axis of a sensor coordinate system and a plurality of Y-line electrodes for detecting the position on the Y axis. Note that the touch sensor 25 may be a self-capacitance-type sensor in which block-shaped electrodes are arranged in a two-dimensional lattice pattern, instead of the mutual capacitance-type sensor described above.

[0022] The electronic pen 14 is a pen-shaped pointing device configured to be capable of one-way or two-way communication with the learner terminal 12. The electronic pen 14 is, for example, an active electrostatic coupling (AES) stylus. A student as a user can write pictures, letters, etc. on the learner terminal 12 by holding the electronic pen 14 and moving the pen tip while pressing it against the touch surface of the learner terminal 12.

[0023] The learning support device 16 is a server computer that performs overall control of learning support, and may be either a cloud-based or on-premise type. Here, the learning support device 16 is illustrated as a single computer, but the learning support device 16 may instead be a group of computers that form a distributed system.

[0024] <Block diagram of learning assistance device 16> Figure 2 is a block diagram showing an example of the configuration of the learning assistance device 16 in Figure 1. Specifically, the learning assistance device 16 includes a communication unit 30, a control unit 32, and a storage unit 34.

[0025] The communication unit 30 is an interface for transmitting and receiving electrical signals to and from an external device, which allows the learning assistance device 16 to provide the question data D1 generated by itself to the learner terminal 12 and to obtain the answer data D2 from the learner terminal 12.

[0026] The control unit 32 is configured with a processor including a CPU or a GPU. The control unit 32 reads and executes the programs and data stored in the storage unit 34, thereby functioning as a data acquisition unit 40, a tag generation unit 42, a relevance estimation unit 44, an exercise processing unit 46, an evaluation processing unit 48, and an output instruction unit 50.

[0027] The data acquisition unit 40 acquires various data (hereinafter also referred to as "deliverable data") as deliverables generated through handwritten input by the user. The data acquisition unit 40 may acquire the deliverable data from an external device via communication, or may acquire the deliverable data by reading it from the storage unit 34.

[0028] Here, "deliverables" refer to digital content created through learning, and may be, for example, letters, illustrations, paintings, or calligraphy. For example, deliverables resulting from exercises correspond to answers to problems. Examples of exercise subjects include Japanese, arithmetic, science, English, and art. Furthermore, deliverable data may include content data, which is a collection of content elements that make up the deliverable, as well as various data related to the creation of the deliverable (hereinafter referred to as "related data").

[0029] The content data may be, for example, ink data (hereinafter referred to as digital ink) for expressing handwritten content. Examples of "ink description languages" for describing digital ink include WILL (Wacom Ink Layer Language), InkML (Ink Markup Language), and ISF (Ink Serialized Format).

[0030] Examples of related data include: [1] user information 62 including user identification information and attributes; [2] "device driver setting conditions" including the resolution, size, and type of the display unit 24, the detection performance and type of the touch sensor 25, and the shape of the pen pressure curve; [3] "drawing application setting conditions" including the type of content, color information for the color palette and brush, and visual effect settings; or [4] "user operation history" stored sequentially throughout the execution of the drawing application.

[0031] The tag generation unit 42 analyzes the stroke data included in the product data acquired by the data acquisition unit 40, and generates a tag information group (hereinafter referred to as a "tag set") that indicates the type of tag that indicates the semantic attribute of a group of strokes and the order in which the tags appear. The detailed function of the tag generation unit 42 will be described in detail with reference to FIG. 3.

[0032] The relevance estimation unit 44 estimates the relevance indicating the degree of relevance between study content items (hereinafter also referred to as "study content items"). The relevance estimation unit 44 may estimate the relevance between study content items based on, for example, [1] graph data or [2] a user's past operation history.

[0033] First, the relevance estimation unit 44 estimates the relevance between learning contents using data for describing a concept graph (i.e., concept graph data 60) or data for describing a knowledge graph (i.e., knowledge graph data 64). For example, the relevance estimation unit 44 may estimate that the relevance is high when the norm between the distributed representation vectors assigned to the learning contents is smaller than a threshold, and that the relevance is low when the norm is larger than the threshold. Alternatively, the relevance estimation unit 44 may estimate that the relevance is high when the learning contents are linked, and that the relevance is low when the learning contents are not linked.

[0034] Second, the relevance estimation unit 44 may estimate that the relevance increases as the number of predetermined operations performed to associate learning content items increases, and may estimate that the relevance decreases as the number of predetermined operations decreases. These "predetermined operations" include (1) a "search" operation to search for learning content items and (2) a "reference" operation to reference learning content items. An example of a search operation is entering the learning content item to be searched for in a search field. An example of a reference operation is selecting a hyperlink related to learning content items.

[0035] The exercise processing unit 46 processes various information related to the exercise (i.e., output learning) performed by the user. The specific configuration of the exercise processing unit 46 will be described in detail with reference to FIG.

[0036] The evaluation processing unit 48 performs various information processing related to the user's evaluation of the exercise. The specific configuration of the evaluation processing unit 48 will be described in detail with reference to FIG.

[0037] The output instruction unit 50 instructs the output device to present information (hereinafter referred to as "learning presentation information") related to learning at least one of the multiple learning contents (hereinafter referred to as "target content") to the user at the timing when the target content is presented. This "output device" may be a device provided in an external device such as the learner terminal 12 (FIG. 1), or may be an output device (not shown) provided in the learning support device 16. The output may be in any form that stimulates the five senses, such as sight, hearing, and touch.

[0038] The above-mentioned learning presentation information may include information for presenting corresponding content that corresponds to the target content as a question or a hint, and for prompting the user to answer the target content by hand. When the target content is a word, examples of the corresponding content include [1] words that have the same concept but different languages ​​(e.g., the relationship between Japanese and English), [2] words that have the same concept but different characters (e.g., the relationship between katakana and kanji, or the relationship between block letters and cursive), [3] a question sentence in which the target content is the correct answer, or [4] a sentence that explains the target content.

[0039] The learning presentation information may also include information that visualizes the relationships between the target content and other learning content (hereinafter referred to as "visualized information"). Examples of representation formats for visualized information include lists, bullet points, correlation diagrams, tree diagrams, graphs, and mind maps. Here, a "mind map" is a representation method that makes it easier to organize memories and generate ideas by illustrating what is in one's head in a form that is close to what is actually in one's mind.

[0040] The visualized information may be displayed in different ways depending on the type of learning. For example, in the case of input learning, the entire visualized information including the target content may be displayed in order to acquire knowledge of the target content. On the other hand, in the case of output learning, the visualized information may be displayed with the target content hidden, and questions may be asked as fill-in-the-blank questions. Alternatively, after output learning of the target content, separately displaying the visualized information including the target content may facilitate systematic knowledge acquisition.

[0041] The storage unit 34 stores programs and data necessary for the control unit 32 to control each component. The storage unit 34 is composed of a non-transitory, computer-readable storage medium. Here, the computer-readable storage medium is composed of [1] a storage device such as an HDD or SSD built into the computer system, and [2] a portable medium such as a magneto-optical disk, ROM, CD-ROM, or flash memory. In the example of FIG. 2, the storage unit 34 stores the question data D1 and answer data D2 (corresponding to "deliverable data"), as well as a shared data group 52 and a personal data group 54.

[0042] The question data D1 includes "question information" regarding questions to be asked during the exercise. This question information includes [1] the subject, format, and time of the exercise, [2] a set of questions and answers, or [3] the number, type, and presentation order of target content. This "target content" corresponds to at least one learning content to be presented. Examples of learning content include words, phrases, phrases, sentences, figures, and illustrations.

[0043] The answer data D2 includes "answer information" regarding answers to questions posed during the exercise. This answer information includes [1] deliverable data generated through the exercise (the content data and related data described above) or [2] meta information regarding the exercise (e.g., whether or not a review operation was performed, the number of reviews, etc.).

[0044] The shared data group 52 corresponds to a data group that can be shared between users when they perform exercises. Specifically, the shared data group 52 includes a database related to questions (hereinafter referred to as a "question DB 56") and a database related to deliverables (hereinafter referred to as a "deliverable DB 58"), and stores conceptual graph data 58.

[0045] The question DB 56 stores available registrations of [1] learning content that constitutes part of a question, [2] learning content that constitutes part of a model answer, or [3] question formats. The deliverable DB 58 stores available registrations of [1] deliverable data generated by multiple users and [2] analysis results of the deliverable data (e.g., statistical data such as average scores and percentage of correct answers).

[0046] The conceptual graph data 60 includes information for describing a conceptual graph. A conceptual graph is a graph (i.e., an ontology graph) that systematically shows the relationships between words and is composed of nodes and links (or edges). Each word that constitutes a node is associated with a coordinate value in an N-dimensional (e.g., N≧3) feature space. In other words, each word is quantified as a "distributed representation" in natural language processing.

[0047] The personal data group 54 corresponds to a data group customized for each individual user performing the exercise. Specifically, the personal data group 54 includes user information 62, knowledge graph data 64, and learning outcome information 66.

[0048] The user information 62 includes various information related to the user. Examples of the user information 62 include: [1] "user identification information" including a name, a nickname, a user ID (Identification), and a password; [2] "attribute information" including a school name, an organization name, a grade, and a gender; and [3] "device identification information" including identification information of the electronic pen 14 (i.e., a pen ID), identification information of the learner terminal 12 (i.e., a terminal ID), and an email address.

[0049] The knowledge graph data 64 includes information for describing a knowledge graph. A knowledge graph is a graph that systematically shows the relationships between words and is composed of nodes and links (or edges). The types of nodes or links may be the same as those in the concept graph data 60, or may be different. Here, the knowledge graph is customized for each user throughout the user's learning process. Specifically, user-specific nodes may be defined and added based on the analysis results of digital ink generated sequentially during learning and non-learning. This allows for the construction of a knowledge graph (i.e., a personalized knowledge graph) that reflects not only the relationships between learning content but also personality, including character, behavioral patterns, and areas of expertise. Note that the knowledge graph may be [1] a simple representation of the relationships between nodes, or [2] a representation that is decorated in a manner that is easy to remember, such as a mind map.

[0050] The learning outcome information 66 includes various information related to the learning outcome. Examples of the learning outcome information 66 include [1] "history information" including the duration of the learning, the number of times the learning was performed, and the total duration of the learning, [2] "score information" including the score and retention level calculated through the exercises, and [3] "timing information" for setting the timing of the presentation of the learning content. The timing information includes, for example, the time when the question was asked, the retention level (peak value) at the time when the question was asked, the forgetting rate, and the question interval.

[0051] <Functional Block Diagram of Tag Generation Unit 42> Fig. 3 is a functional block diagram of the tag generation unit 42 shown in Fig. 2. Specifically, the tag generation unit 42 includes an attribute estimator 70, a switch 72, a graphic identifier 74, a text identifier 75, a mathematical expression identifier 76, and a tag integrator 78.

[0052] The attribute estimator 70 is a learning device that receives stroke data representing a collection of strokes and outputs semantic attributes of the group of strokes. Attribute estimation is performed, for example, using the method disclosed in International Publication No. 2020 / 090356. Types of attributes include, for example, graphic, text, mathematical formula, chemical formula, and non-applicable.

[0053] The switch 72 has one input terminal and three output terminals, and switches the output destination depending on the attribute supplied from the attribute estimator 70. Specifically, if the attribute is "graphic", the stroke data is supplied to the input side of a graphic classifier 74 via the switch 72. If the attribute is "text", the stroke data is supplied to the input side of a text classifier 75 via the switch 72. If the attribute is "mathematical formula", the stroke data is supplied to the input side of a mathematical formula classifier 76 via the switch 72.

[0054] The figure classifier 74 is a learning device that receives stroke data as input and outputs a figure tag indicating the identification information of the figure. Various artificial intelligence techniques, including hierarchical neural networks and convolutional neural networks, may be used to identify the figure. The figure tag may be, for example, a figure element such as a circle, a triangulation scale, or a line segment, or an object such as a dog, a cat, a tree, or a pencil.

[0055] The text classifier 75 is a learning device that receives stroke data as input and outputs text tags that indicate text identification information. Various artificial intelligence techniques, including hierarchical neural networks and convolutional neural networks, may be used to identify the text. The text tag may be, for example, a character such as "A," "B," or "C," a word or sentence consisting of multiple characters, or a language such as Japanese, English, or Chinese.

[0056] The mathematical formula identifier 76 is a learning device that receives stroke data as input and outputs a text tag indicating the identification information of the mathematical formula. Various artificial intelligence techniques, including hierarchical neural networks and convolutional neural networks, may be used to identify the mathematical formula. The mathematical formula tag may be, for example, letters such as "X" and "Y," numbers such as "1," "2," and "3," operators such as "+" and "=", or types of formulas such as theorem, transformation, and final solution.

[0057] The tag integrator 78 sequentially integrates the "graphic tags" from the graphic identifier 74, the "text tags" from the text identifier 75, and the "mathematical formula tags" from the mathematical formula identifier 76, and outputs a tag set in association with the order in which the tags appear. If there is some relationship between consecutive tags, they may be output as graph structure data with each tag as a node.

[0058] The calculation rules of the tag generation unit 42 may be uniform regardless of the user, or may be customized for each user. In the latter case, it is desirable that a group of learning parameters defining the calculation rules of the tag generation unit 42 be pre-stored in association with identification information that can identify the user, and be read out at appropriate times using the identification information included in the deliverable data. Examples of the identification information include an account for using the learning support service, various IDs related to the electronic pen 14, or the learner terminal 12.

[0059] <Functional Blocks of Exercise Processing Unit 46> Fig. 4 is a functional block diagram of the exercise processing unit 46 shown in Fig. 2. Specifically, the exercise processing unit 46 includes a request receiving unit 80, a mastery calculation unit 82, a timing setting unit 84, and a question determination unit 86.

[0060] The request receiving unit 80 receives a signal from the user requesting the start of an exercise (i.e., an exercise start request). This exercise start request includes either [1] information that can identify the user (i.e., user information 62) or [2] “format information” regarding the exercise format, including the subject, format, and time of the exercise.

[0061] The retention calculation unit 82 calculates the retention of the user's knowledge of the learning content. After identifying the relevant user from the user information 62 provided by the request receiving unit 80, the retention calculation unit 82 calculates the retention of each learning content by referencing the learning outcome information 66 for each user. This "retention" is defined, for example, so that a larger value indicates a state in which the knowledge is more firmly established (or memorized), while a smaller value indicates a state in which the knowledge is less firmly established (or memorized). Specifically, the retention may be defined as a score out of 100 (i.e., within the range of 0 to 100 points).

[0062] After the retention level of the learning content temporarily increases through practice, the retention level calculation unit 82 may [1] calculate the retention level for each learning content as a constant regardless of the time elapsed since the most recent learning point, or [2] calculate the retention level for each learning content according to a mathematical model that shows changes over time since the learning point. Examples of this "learning point" include [1] the time when output learning is performed, [2] the time when input learning is performed, or [3] the time when handwritten input is performed on an occasion other than learning. Furthermore, an example of a "mathematical model" is the Evins-House forgetting model. This "Evins-House forgetting model" corresponds to a model that shows the forgetting characteristics of human medium-term memory or long-term memory.

[0063] The timing setting unit 84 sets the presentation timing of the learning content included in the knowledge graph data 64 according to the level of retention calculated by the level of retention calculation unit 82. For example, the timing setting unit 84 may set the presentation timing so that [1] the frequency of presentation (or learning frequency) increases as the level of retention decreases, or [2] the frequency of presentation (or learning frequency) decreases as the level of retention increases. Alternatively, the timing setting unit 84 may set the presentation timing so that [1] the time interval between presentations (or question intervals) decreases as the level of retention decreases, or [2] the time interval between presentations (or question intervals) increases as the level of retention increases.

[0064] Furthermore, the timing setting unit 84 may set the presentation timing for not only one target content but also two or more target contents. When the fixation level of the first target content is defined as the first fixation level and the fixation level of the second target content is defined as the second fixation level, the presentation timing may be determined as follows. For example, [1] when the second fixation level is lower than the first fixation level, the first target content and the second target content may be presented simultaneously or at similar time intervals. Alternatively, [2] when the second fixation level is higher than the first fixation level, the first target content and the second target content may be presented simultaneously or at similar time intervals. Alternatively, [3] when the second fixation level is equal to the first fixation level within an acceptable range, the first target content and the second target content may be presented simultaneously or at similar time intervals.

[0065] An example of simultaneously presenting target contents is when a first target content is set as a "question" and a second target content is presented as a "hint" in a fill-in-the-blanks format.An example of presenting target contents at close time intervals is when a so-called "knock-type" format questions are used and both the first target content and the second target content are presented sequentially as "questions" within one practice session.

[0066] The question determination unit 86 determines the content of questions to be asked in the exercise in accordance with the presentation timing set by the timing setting unit 84. The content of questions includes [1] the subject, format, and time of the exercise, [2] a set of questions and answers, or [3] the number, type, and presentation order of target content. This determination refers to the question DB 56 or knowledge graph data 64 as necessary. Through this determination process, question data D1 is generated.

[0067] <Functional Blocks of Evaluation Processing Unit 48> Fig. 5 is a detailed functional block diagram of the evaluation processing unit 48 shown in Fig. 2. Specifically, the evaluation processing unit 48 is configured to include an index generation unit 90, a score calculation unit 92, and a fixation calculation unit 94.

[0068] The index generating unit 90 performs data processing on the answer data D2 corresponding to the question data D1 to generate one or more indexes (hereinafter also referred to as "index set") for evaluating the learning outcome. The index generating unit 90 also performs data processing on the model data D3 indicating the model answer to the question to generate an index set indicating the evaluation criteria. Hereinafter, the index generated from the answer data D2 will be referred to as the "evaluation target value," and the index generated from the model data D3 will be referred to as the "ideal value."

[0069] Prior to generating an index set, the index generating unit 90 calculates a collection of features (hereinafter referred to as a "feature set"). Examples of the features include [1] a "success / failure feature" relating to the success or failure of the content of handwritten input by the user, [2] an "accuracy feature" relating to the accuracy of the handwritten input, or [3] a "smoothness feature" relating to the smoothness of the handwritten input.

[0070] Examples of success / failure features include a success / failure flag indicating correct / incorrect, the presence or absence of an answer, the presence or absence of a hint, or the level of incorrectness. Examples of accuracy features include the accuracy of characters or the accuracy of stroke shapes. Examples of smoothness features include the change in writing pressure over time, the time interval between stroke operations, or the number of undo operations (or the number of rewrites).

[0071] The score calculation unit 92 calculates a score for each piece of learning content presented or quizzed using the index set generated by the index generation unit 90. The score calculation unit 92 calculates a score that indicates, for example, the degree of agreement between the evaluation target value and the ideal value. In this case, the closer the evaluation target value is to the ideal value, the higher the score, and the farther the evaluation target value is from the ideal value, the lower the score. The score may be defined in the same way as the degree of retention (e.g., a perfect score of 100) described above, or may be defined differently from the degree of retention.

[0072] The retention calculation unit 94 calculates the user's knowledge retention for each learning content based at least on the score calculated by the score calculation unit 92. This retention is calculated, for example, by adding the learning effect amount to the original score. The learning effect amount is calculated so that it increases the following: (1) the current score is lower; (2) the number of reviews is greater; or (3) the time elapsed since the previous presentation or learning point is longer.

[0073] 6 is a diagram showing an example of a network structure of a learning device 96 included in the index generating unit 90 of FIG. The learning device 96 is configured, for example, as a hierarchical neural network, and includes an input layer LI, an intermediate layer LM, and an output layer LO. The computation rules of the learning device 96 are determined by the values ​​of learning parameters constituting a parameter group. The learning parameters may include, for example, coefficients describing the activation functions of the computation units, weighting coefficients corresponding to the strength of synaptic connections, the number of computation units constituting each layer, the number of intermediate layers LM, etc.

[0074] The input layer LI is a layer that inputs the success / failure feature set, the accuracy feature set, and the smoothness feature set, and is composed of a plurality of (for example, N) arithmetic units.

[0075] The hidden layer LM is configured with one or more layers and has a dimension reduction function for reducing the number of dimensions of the input vector. In other words, it is desirable that the number of operation units constituting the hidden layer LM is less than N.

[0076] The output layer LO is a layer that outputs an index set, which is one or more indexes, and in the example shown in the figure, is composed of n (1≦n<N) arithmetic units.

[0077] [Operation of learning assistance system 10] The learning assistance system 10 incorporating the learning assistance device 16 in this embodiment is configured as described above. Next, the operation of the learning assistance system 10 will be described with reference to Figures 7 to 24 in addition to Figures 1 to 6.

[0078] <1. Knowledge Graph Update Operation> First, the knowledge graph update operation by the learning assistance system 10 will be described with reference to the sequence diagram of FIG. 7 and FIGS. 8 to 11.

[0079] In step SP10 of FIG. 7, the learner terminal 12 accepts the learner's handwritten input operation and sequentially generates product data including stroke data.

[0080] 8 is a diagram showing an example of a collection G0 of strokes displayed on the display unit 24 of the learner terminal 12. Here, the collection G0 represents a handwritten shopping list, consisting of the items "milk," "bread," "eggs," and "apples" from top to bottom. In this case, the collection G0 includes a group G1, and the group G1 is made up of four groups G2 to G5.

[0081] 9 is a diagram showing an example of the data structure of digital ink 100. Digital ink 100 has a data structure in which [1] document metadata, [2] ink semantics, [3] device data, [4] stroke data, [5] group data, and [6] context data are sequentially arranged. When digital ink 100 is generated, the document metadata, device data, stroke data, and context data have already been determined, but the semantic data and group data have not yet been determined. In other words, strokes have not yet been classified or assigned semantic attributes.

[0082] The stroke data 102 is data for describing individual strokes that make up a handwritten work, and indicates the shape of the strokes that make up the work and the order in which the strokes are written. As can be seen from FIG. 9 , one stroke is described by multiple point data items arranged sequentially within the <trace> tag. Each point data item consists of at least a designated position (X coordinate, Y coordinate) and is separated by a delimiter such as a comma. For convenience of illustration, only point data indicating the start and end points of the stroke are shown, and point data indicating multiple intermediate points are omitted. In addition to the designated positions described above, this point data may also include the order in which the strokes were created or edited, the amount of pressure applied by the electronic pen 14, and the attitude (e.g., tilt or orientation) of the electronic pen 14.

[0083] In step SP12 of FIG. 7, the learner terminal 12 transmits the product data accumulated through the repeated execution of step SP10 to the learning assistance device 16.

[0084] In step SP14, the learning support device 16 (more specifically, the data acquisition unit 40) receives and acquires the deliverable data transmitted in step SP12.

[0085] In step SP16, the learning support device 16 (more specifically, the tag generation unit 42) analyzes the deliverable data received in step SP14 and generates a tag set.

[0086] 10 is a diagram showing the results of assigning semantic attributes to a stroke collection G0. The type of group G1 is "LIST" and the value is "NULL". The type of group G2 is "LIST_ITEM" and the value is "milk". The type of group G3 is "LIST_ITEM" and the value is "bread". The type of group G4 is "LIST_ITEM" and the value is "eggs". The type of group G5 is "LIST_ITEM" and the value is "apples".

[0087] 7, the learning assistance device 16 (more specifically, the relevance estimation unit 44) checks whether or not to update the knowledge graph corresponding to the learner. If the update timing has arrived, the learning assistance device 16 proceeds to the next step, SP20.

[0088] In step SP20, the relevance estimation unit 44 of the learning support device 16 updates the knowledge graph by editing the knowledge graph data 64.

[0089] 11 is a diagram showing an example of the data structure of the knowledge graph data 64. This knowledge graph data 64 is, for example, data in a table format showing the correspondence between information about nodes (i.e., node information) and information about links (i.e., link information). The node information includes [1] identification information of the first node (first node ID), [2] the type of the node, and [3] the name of the node. The link information includes [1] identification information of the second node having a connection relationship with the first node (second node ID) and [2] the type of the relationship attribute.

[0090] Thereafter, the learning assistance system 10 repeatedly executes steps SP10 to SP20 in FIG. 7, thereby successively updating the knowledge graph specific to the learner.

[0091] 2. Learning Assistance Operation Next, an overview of the learning assistance operation by the learning assistance system 10 will be described with reference to the sequence diagram of FIG. 12 and FIGS. 13 to 16. FIG.

[0092] In step SP30 of FIG. 12, the learner terminal 12 accepts a predetermined operation for requesting the start of the exercise (that is, a start request operation).

[0093] In step SP32, the learner terminal 12, in response to the start request operation performed in step SP30, transmits a signal requesting the start of the exercise (i.e., a start request signal) to the learning assistance device 16.

[0094] In step SP34, the learning assistance device 16 receives the start request signal sent in step SP32 from the learner terminal 12, and thereby acquires the user information 62 corresponding to the learner.

[0095] In step SP36, the learning support device 16 (more specifically, the exercise processing unit 46) uses the user information 62 acquired in step SP34 to generate question data D1 according to the learner's level of knowledge retention.

[0096] In step SP38, the learning support device 16 (more specifically, the output instruction unit 50) transmits the question data D1 generated in step SP36 to the corresponding learner terminal 12.

[0097] In step SP40, the learner terminal 12 receives the question data D1 transmitted in step SP38, generates a display signal showing the exercise screens 110, 120, and 130 using the question data D1, and then supplies the display signal to the display unit 24. As a result, the exercise screens 110, 120, and 130 are displayed on the display unit 24.

[0098] 13 is a diagram showing a first example of an exercise screen 110 displayed on the display unit 24 of the learner terminal 12. This exercise screen 110 corresponds to "learning presentation information" presented to the learner for English vocabulary dictation practice. Specifically, the exercise screen 110 includes a question field 112 showing the content of the question, an answer field 114 for handwritten input of an answer, and a button 116 labeled "Next." The learner handwrites the English word corresponding to "Wednesday" presented in the question field 112 into the answer field 114 (the correct answer is "Wednesday"), and then performs a touch or tap operation to select the "Next" button 116. This displays the correct answer or moves to the next question.

[0099] FIG. 14 illustrates a second example of an exercise screen 120 displayed on the display unit 24 of the learner terminal 12. This exercise screen 120 corresponds to "learning presentation information" presented to the learner for English vocabulary dictation practice. Specifically, the exercise screen 120 includes hint information fields 122 and 123 that display hints for the question, a question field 124 that suggests the question, an answer field 126 for handwritten input of an answer, and a button 128 labeled "Next." The two hint information fields 122 and 123 display a list of English words representing the days of the week. The learner handwrites the English word that should be entered in the question field 124 (the correct answer is "Wednesday") in the answer field 126, and then touches or taps the "Next" button 128. This displays the correct answer or moves to the next question.

[0100] FIG. 15 illustrates a third example of an exercise screen 130 displayed on the display unit 24 of the learner terminal 12. This exercise screen 130 corresponds to "learning presentation information" presented to the learner for English vocabulary dictation practice. Specifically, the exercise screen 130 includes a question field 132 displaying a mind map as a question, an answer field 134 for handwritten input of an answer, and a button labeled "Next" 136. This mind map depicts a central word (i.e., target content) connected to other words 140 via connecting lines 142 that mimic synaptic connections. The learner handwrites an English word to be filled in the blank 138 in the answer field 134, and then touches or taps the "Next" button 136. This displays the correct answer or moves to the next question.

[0101] In step SP44 of FIG. 12, the learner terminal 12 accepts a predetermined operation for requesting the end of the exercise (that is, an end request operation).

[0102] In step SP46, the learner terminal 12 transmits an end request signal including the accumulated answer data D2 to the learning assistance device 16 at the time when the end request operation is received in step SP42.

[0103] In step SP48, the learning assistance device 16 receives the end request signal sent in step SP46 from the learner terminal 12, and thereby acquires the answer data D2 corresponding to the question data D1.

[0104] In step SP50, the evaluation processing unit 48 (more specifically, the index generation unit 90) of the learning support device 16 performs data processing on the answer data D2 acquired in step SP46 to generate an index set to be used for evaluation.

[0105] In step SP52, the evaluation processing unit 48 (more specifically, the score calculation unit 92) calculates a score for each of the presented learning content (i.e., target content) based on the index set generated in step SP50.

[0106] In step SP54, the evaluation processing unit 48 (more specifically, the retention calculation unit 94) calculates the retention of the learner's knowledge based on the score calculated in step SP52.

[0107] In step SP56, the evaluation processing unit 48 generates learning outcome information 66 including the degree of retention calculated in step SP54. This learning outcome information 66 is stored in the storage unit 34 of the learning support device 16 in a state associated with the learner.

[0108] 16 is a diagram showing a first example of the data structure of the learning outcome information 66 in FIG. 2. The learning outcome information 66 is data in a table format that shows the correspondence between, for example, [1] the identification number (content ID) of the learning content, [2] the degree of understanding, [3] the question interval, and [4] the question history. "Degree of understanding" refers to the degree of understanding at the time of the most recent question (i.e., the value calculated by the degree of understanding calculation unit 94). The question history includes pairs of the time of question presentation and the score.

[0109] Thereafter, the learning assistance system 10 can support the learner in repeated learning by repeatedly executing steps SP30 to SP56 in FIG.

[0110] 3. Operation of Generating Question Data (First Example) Next, a first example of the operation of generating question data D1 by learning support device 16 (step SP36 in FIG. 12) will be described in detail with reference to the flowchart in FIG. 17 and FIGS. 18 to 20. In this first example, it is assumed that learning support device 16 generates question data D1 using learning outcome information 66 in FIG. 16.

[0111] 17, the exercise processing unit 46 of the learning support device 16 (more specifically, the request receiving unit 80 of FIG. 4) checks whether or not it has received an exercise start request from the learner terminal 12. If the exercise start request has not yet been received (step SP60: NO), the exercise processing unit 46 remains in step SP60 until it receives the exercise start request. If the exercise start request has been received (step SP60: YES), the exercise processing unit 46 proceeds to the next step SP62.

[0112] In step SP62, the practice processing unit 46 (more specifically, the timing setting unit 84) sets the timing for presenting the learning content. Prior to this setting, the timing setting unit 84 references the learning outcome information 66 (FIG. 16) corresponding to the learner and obtains the question interval for each learning content.

[0113] Figure 18 shows the Evins-House forgetting curve. The horizontal axis of the graph represents elapsed time, and the vertical axis of the graph represents memory retention. As can be seen from this graph, memory retention monotonically declines as time passes from the time of memorization. For example, if the memory retention rate at the time of memorization is 1, then after one hour, the memory retention rate drops sharply to 0.44. Then, as time passes, the rate of decline gradually slows, and after one month, the memory retention rate drops to 0.21.

[0114] FIG. 19 is a diagram showing an example of the process of knowledge consolidation. The horizontal axis of the graph indicates elapsed time (unit: days), and the vertical axis of the graph indicates the degree of consolidation (or score). Elapsed time = 0 days corresponds to the time of the first practice. As can be seen from FIG. 19, knowledge can be more efficiently consolidated by repeating the learning process for one learning content, for example, by setting [1] 2 days later (time interval = 2 days), [2] 7 days later (time interval = 5 days), or [3] 30 days later (time interval = 23 days).

[0115] FIG. 20 shows a first example of a method for setting presentation timing. The axis of the graph represents the passage of time. Here, we assume that the presentation timing for three learning content items C1, C2, and C3 is being set. Each arrow indicates the question suspension period for learning content items C1 to C3. The start point of the question suspension period corresponds to the most recent question being asked. The end point of the question suspension period is the point in time when the question interval has elapsed from the current question being asked. In the example of FIG. 19, at the time of the next question being asked, learning content items C1 and C2, which have already passed the question suspension period, are set to "available for questioning." In contrast, at the time of the next question being asked, learning content item C3, which has not yet passed the question suspension period, is set to "unavailable for questioning."

[0116] In step SP64 of FIG. 17, the exercise processing unit 46 (more specifically, the question determination unit 86) determines the question content (specifically, the type of target content and the presentation order) according to the presentation timing set in step SP64.

[0117] In this way, the exercise processing unit 46 of the learning support device 16 completes the generation operation of step SP36 in Fig. 17. As a result, an exercise is performed that takes into account memory phases based on neuroscientific knowledge.

[0118] 4. Operation of Generating Question Data (Second Example) Next, a second example of the operation of generating question data D1 by the learning assistance device 16 will be described in detail with reference to FIGS. 21 to 24. FIG.

[0119] 21 is a diagram showing a second example of the data structure of the learning outcome information 66A in FIG. 2. The learning outcome information 66A is data in a table format showing the correspondence between, for example, [1] the learning content identification number (content ID), [2] the retention peak value, [3] the forgetting rate, and [4] the question history. The "retention peak value" refers to the retention rate at the most recent practice session. The "forgetting rate" refers to the exponential multiplier in the forgetting curve; the larger the value, the faster the forgetting rate.

[0120] 22 is a second flowchart of the operation for generating the question data D1 (step SP36A in FIG. 12). In this second example, it is assumed that the learning support device 16 generates the question data D1 using the learning outcome information 66A in FIG.

[0121] In step SP70, the exercise processing unit 46 of the learning support device 16 (more specifically, the request receiving unit 80 in FIG. 4) checks whether or not it has received an exercise start request from the learner terminal 12. If the exercise start request has not yet been received (step SP70: NO), the exercise processing unit 46 remains in step SP70 until it receives the exercise start request. If the exercise processing unit 46 has received an exercise start request (step SP70: YES), it proceeds to the next step SP72.

[0122] In step SP72, the practice processing unit 46 (more specifically, the retention calculation unit 82) calculates the retention of knowledge for each learning content. In the second example, a retention level that reflects changes over time due to forgetting is used.

[0123] 23 is a diagram showing an example of the change in retention rate over time based on a forgetting model. The horizontal axis of the graph represents time (unit: days), and the vertical axis of the graph represents retention rate (or score). A virtual curve 150 includes a curved portion 152 connecting two points P0 and P1, a straight line portion 154 connecting two points P1 and P2, and a curved portion 156 passing through two points P2 and P3. Both curved portions 152 and 156 are represented by approximate curves that decrease exponentially.

[0124] When the length (time width) of the section in the curved portion 152 is defined as T1 and the forgetting rates of the curved portions 152 and 156 are defined as γ1 and γ2, respectively, γ2 can be calculated, for example, by the following formula (1): γ2=(1-α)(1+βT1)γ1 (1)

[0125] Here, α (0<α<1) is a coefficient that indicates the efficiency of repeated learning. β (β>0) is a coefficient that indicates the efficiency of retention. For example, by increasing the question interval T1, the forgetting rate is updated to a high value, and the retention rate drops sharply over time. Conversely, by shortening the question interval T1 and repeating learning, the forgetting rate gradually drops, and the retention rate can be maintained at a high level.

[0126] In step SP74 of FIG. 22, the practice processing unit 46 (more specifically, the timing setting unit 84) sets the timing for presenting the study content based on the level of understanding calculated in step SP72.

[0127] Figure 24 shows a second example of a presentation timing setting method. The axis of the graph indicates the level of understanding. Here, we assume that learning content C0 is presented, and set the presentation timing for four learning content items C1, C2, C3, and C4. Each dot indicates the understanding level of learning content items C1 to C4. The understanding levels R0 to R4 satisfy the magnitude relationship R1 < R2 < R3 < R0 < R4. Circles (C0, C2 to C4) indicate high relevance between the learning content items. In contrast, triangles (C1) indicate low relevance with other learning content items.

[0128] For example, at the time of the next question, learning content C2, which (1) has a high relevance to learning content C0 and (2) has the lowest level of retention, may be set as "available for questioning." This can improve retention of learning content C0 and improve learning efficiency. Note that if relevance is not a consideration, learning content C1 may be selected instead of learning content C2.

[0129] For example, at the time of the next question, learning content C4 that (1) has a high relevance to learning content C0 and (2) has the highest level of retention may be set as "available for questioning." By presenting learning content C4 with a high level of retention, the learner is more likely to get the correct answer, which increases the learner's motivation to study.

[0130] For example, at the time of the next question, the learning content C3 that (1) has a high relevance to the learning content C0 and (2) has a level of mastery closest to R0 may be set as "available for questioning." By presenting multiple learning content C4s with similar levels of mastery together, the learner can systematically acquire knowledge of the same level of difficulty, thereby improving learning efficiency.

[0131] In step SP76 of FIG. 22, the exercise processing unit 46 (more specifically, the question determination unit 86) determines the question content (specifically, the type of target content and the presentation order) according to the presentation timing set in step SP74.

[0132] In this way, the exercise processing unit 46 of the learning support device 16 completes the generation operation of step SP36A in Fig. 22. As a result, an exercise is performed that takes into account memory phases based on neuroscientific knowledge.

[0133] [Summary of the embodiment] As described above, the learning support system 10 in this embodiment includes a user terminal (here, the learner terminal 12) that generates deliverable data, which is data indicating a deliverable obtained through handwritten input by a user and includes stroke data 102 for describing strokes.

[0134] The learning support device 16 includes an output instruction unit 50 that instructs an output device (here, the display unit 24) to present learning presentation information related to learning of a target content, which is at least one learning content among a plurality of learning contents, to a user at the presentation timing of the target content; a data acquisition unit 40 that acquires result data (here, answer data D2) that indicates the result obtained by the user's handwritten input to the learning presentation information presented using the display unit 24; a fixation calculation unit 82, 94 that performs an analytical process on the answer data D2 to calculate the fixation level of the user's knowledge regarding the target content; and a timing setting unit 84 that sets the next presentation timing of the target content according to the calculated fixation level.

[0135] Furthermore, according to the learning support method and learning support program of this embodiment, one or more computers (or one or more processors) execute an instruction step (SP38) of instructing an output device (here, the display unit 24) to present learning presentation information to the user at the timing of presenting the target content, an acquisition step (SP48) of acquiring product data (i.e., answer data D2) obtained by the user's handwritten input to the learning presentation information, a calculation step (SP54, SP72) of performing an analysis process on the answer data D2 to calculate the user's level of knowledge retention regarding the target content, and a setting step (SP62, SP74) of setting the next presentation timing of the target content based on the calculated level of retention.

[0136] In this way, by setting the next presentation timing of the target content according to the degree of knowledge calculated from the answer data D2 obtained by the user's handwriting, it is possible to present the learning content to the user in a timely manner, thereby further improving the degree of knowledge retention through a learning style in which multiple learning content items are presented to the user.

[0137] Furthermore, the timing setting unit 84 may set the next presentation timing so that the frequency of presentation increases as the fixation level decreases, or so that the frequency of presentation decreases as the fixation level increases.

[0138] The retention calculation unit 82 may also calculate the retention of each learning content item according to a mathematical model that indicates changes over time from the time the learning content was presented or studied. In particular, the Evins-House forgetting model may be used as the mathematical model. This enables repeated learning that takes into account memory phases based on neuroscience findings, allowing knowledge to be retained more efficiently.

[0139] The degree of fixation may be calculated using an accuracy feature related to the accuracy of the user's handwriting input or a smoothness feature related to the smoothness of the handwriting input. The smoothness feature may include at least one of a change in writing pressure over time, a time interval between stroke operations, and the number of undo operations.

[0140] Furthermore, when the fixation level of the first target content calculated by the fixation level calculation units 82, 94 is defined as the first fixation level, and the fixation level of the second target content is defined as the second fixation level, the timing setting unit 84 may set the next presentation timing so that the first target content and the second target content are presented simultaneously or at a similar time interval when the second fixation level is lower than the first fixation level. This makes it possible to raise the level of fixation of the second target content and improve learning efficiency.

[0141] Alternatively, if the second fixation level is higher than the first fixation level, the timing setting unit 84 may set the next presentation timing so that the first target content and the second target content are presented simultaneously or at a similar time interval. By deliberately presenting learning content with a high fixation level, the learner is more likely to get the correct answer, which can increase the learner's motivation to study.

[0142] Alternatively, the timing setting unit 84 may set the next presentation timing so that the first target content and the second target content are presented simultaneously or at similar time intervals when the second level of fixation is equal to the first level of fixation within an acceptable range. By presenting multiple learning contents with similar levels of fixation together, the learner can systematically acquire knowledge of the same level of difficulty, thereby improving learning efficiency.

[0143] Furthermore, when there are two or more target candidates that are candidates for the second target content, the timing setting unit 84 may select a target candidate that has a relatively high relevance to the first target content as the second target content. By presenting not only the first target content but also the second target content, which is related knowledge to the first target content, knowledge retention is more likely to be achieved.

[0144] Furthermore, if the learning assistance device 16 further includes a relevance estimation unit 44 that estimates the degree of relevance between learning contents, the timing setting unit 84 may set the next presentation timing so that two or more learning contents that are estimated to have a high relevance by the relevance estimation unit 44 are presented simultaneously or at close time intervals. By presenting not only relevant knowledge but also related knowledge, knowledge retention is more likely to be achieved.

[0145] The relevance estimation unit 44 may also estimate the relevance using a knowledge graph that shows the relationships between learning content items. In particular, the knowledge graph may be generated or updated based on the user's past operation history.

[0146] The relevance estimation unit 44 may also estimate that the greater the number of predetermined operations performed to associate learning content with each other, the higher the relevance, and the fewer the number of predetermined operations performed, the lower the relevance. In particular, the predetermined operations may be operations for searching for or referencing learning content.

[0147] The learning presentation information may also include: [1] information that presents corresponding content that corresponds to the target content as a question or hint, allowing the user to answer the target content by hand; [2] information that presents related content that is related to the target content as a question or hint, allowing the user to answer the target content by hand; or [3] information that visualizes the relationship between the target content and other learning content.

[0148] [Modifications] The present invention is not limited to the above-described embodiment, and can be freely modified without departing from the spirit of the present invention. Alternatively, the respective configurations may be arbitrarily combined within the scope of no technical contradiction. Alternatively, the execution or order of each step constituting the flowchart may be changed within the scope of no technical contradiction.

[0149] In the above embodiment, the dictation exercises for English words (FIGS. 13 to 15) are used as an example, but the types of exercises are not limited to this. For example, the type of exercise may be a "transcription exercise" in which model learning content is transcribed, or a "free writing exercise" in which given questions are freely written.

[0150] In the above embodiment, the pen-shaped electronic pen 14 has been described as an example, but the form of the position indicator is not limited to this pen shape. Also, while the electronic pen 14 has been described as an AES-type stylus, the electronic pen 14 may alternatively be an electromagnetic induction (EMR)-type stylus. In this device configuration, the learner terminal 12 is provided with a planar sensor (or digitizer) having multiple loop coils formed thereon, and the electronic pen 14 is provided with a receiving circuit for receiving magnetic field signals emitted by the planar sensor.

[0151] [Explanation of symbols] 10... learning support system, 12... learner terminal (user terminal), 14... electronic pen, 16... learning support device, 40... data acquisition unit, 42... relevance estimation unit, 50... output instruction unit, 82, 94... fixation calculation unit, 84... timing setting unit, 100... digital ink, 102... stroke data, 110, 120, 130... exercise screen (learning presentation information), D1... question data, D2... answer data (deliverable data)

Claims

1. A learning support device comprising: an output instruction unit that instructs an output device to present learning presentation information related to learning of a target content, which is at least one learning content among a plurality of learning contents, to a user at the timing of presentation of the target content; a data acquisition unit that acquires product data indicating a product obtained by the user's handwriting input on the learning presentation information presented using the output device, the product data including stroke data for describing strokes; a grasp calculation unit that performs an analytical process on the product data acquired by the data acquisition unit to calculate the grasp of the user's knowledge of the target content; and a timing setting unit that sets the next presentation timing of the target content in accordance with the grasp calculated by the grasp calculation unit.

2. The learning support device according to claim 1, wherein the timing setting unit sets the next presentation timing so that the frequency of presentation increases as the fixation level decreases, or so that the frequency of presentation decreases as the fixation level increases.

3. The learning support device according to claim 2, wherein the retention calculation unit calculates the retention of each of the learning content items according to a mathematical model that indicates changes over time from the time the learning content items are presented or studied.

4. The learning support device according to claim 3, wherein the mathematical model is the Evins-House forgetting model.

5. The learning support device of claim 1, wherein the degree of fixation is calculated using a success / failure feature relating to the success or failure of the content of the handwritten input by the user, an accuracy feature relating to the accuracy of the handwritten input, or a smoothness feature relating to the smoothness of the handwritten input.

6. The learning support device according to claim 5, wherein the smoothness feature amount includes at least one of a change in writing pressure over time, a time interval between stroke operations, and a number of undo operations.

7. The learning support device of claim 1, wherein, when the fixation level of the first target content calculated by the fixation level calculation unit is defined as a first fixation level and the fixation level of the second target content is defined as a second fixation level, the timing setting unit sets the next presentation timing so as to present the first target content and the second target content simultaneously or at a similar time interval when the second fixation level is lower than the first fixation level.

8. The learning support device of claim 1, wherein, when the fixation level of the first target content calculated by the fixation level calculation unit is defined as a first fixation level and the fixation level of the second target content is defined as a second fixation level, the timing setting unit sets the next presentation timing so as to present the first target content and the second target content simultaneously or at a similar time interval when the second fixation level is higher than the first fixation level.

9. The learning support device of claim 1, wherein, when the fixation level of the first target content calculated by the fixation level calculation unit is defined as a first fixation level and the fixation level of the second target content is defined as a second fixation level, the timing setting unit sets the next presentation timing so as to present the first target content and the second target content simultaneously or at a similar time interval when the second fixation level is equal to the first fixation level within an acceptable range.

10. A learning support device as described in any one of claims 7 to 9, wherein when there are two or more target candidates that are candidates for the second target content, the timing setting unit selects the target candidate that has a relatively high relevance to the first target content as the second target content.

11. The learning support device of claim 1, further comprising a relevance estimation unit that estimates a relevance indicating the degree of relevance between the learning contents, and the timing setting unit sets the next presentation timing so as to present two or more learning contents estimated by the relevance estimation unit to have a high relevance simultaneously or at close time intervals.

12. The learning support device according to claim 11, wherein the relevance estimation unit estimates the relevance using a knowledge graph indicating the relationships between the learning contents.

13. The learning support device according to claim 12, wherein the knowledge graph is generated or updated based on a past operation history by the user.

14. The learning support device of claim 11, wherein the relevance estimation unit estimates that the degree of relevance is higher as the number of specified operations performed to associate the learning contents with each other increases, and estimates that the degree of relevance is lower as the number of specified operations decreases.

15. The learning support device according to claim 14, wherein the predetermined operation is an operation for searching or referencing the learning content.

16. The learning support device according to claim 1, wherein the learning presentation information includes information for presenting corresponding content that corresponds to the target content as a question or a hint, and for allowing the user to answer the target content by handwriting.

17. The learning support device according to claim 1, wherein the learning presentation information includes information for presenting related content related to the target content as a question or a hint, and for allowing the user to answer the target content by handwriting.

18. The learning support device according to claim 1, wherein the learning presentation information includes information that visualizes the relationship between the target content and other learning content.

19. A learning support program that causes one or more computers to execute the following steps: an instruction step of instructing an output device to present learning presentation information related to learning of a target content, which is at least one learning content among multiple learning contents, to a user at the timing of presentation of the target content; an acquisition step of acquiring product data, which is data indicating a product obtained by the user's handwriting input on the learning presentation information presented using the output device, and which includes stroke data for describing strokes; a calculation step of performing an analysis process on the acquired product data to calculate the user's level of knowledge retention regarding the target content; and a setting step of setting the next presentation timing of the target content depending on the calculated level of retention.

20. A learning support method in which one or more computers execute the following steps: an instruction step of instructing an output device to present learning presentation information related to learning of a target content, which is at least one learning content among a plurality of learning contents, to a user at the timing of presentation of the target content; an acquisition step of acquiring product data indicating a product obtained by the user's handwritten input on the learning presentation information presented using the output device, the product data including stroke data for describing strokes; a calculation step of performing an analysis process on the acquired product data to calculate the user's degree of knowledge retention regarding the target content; and a setting step of setting the next presentation timing of the target content according to the calculated degree of retention.

21. A learning support system comprising: a user terminal that generates product data indicating a product obtained by a user's handwritten input, the product data including stroke data for describing strokes; and a learning support device configured to communicate with the user terminal, wherein the learning support device comprises: an output instruction unit that instructs an output device to present learning presentation information regarding learning of a target content, which is at least one learning content among a plurality of learning contents, to the user at the presentation timing of the target content; a data acquisition unit that acquires from the user terminal the product data corresponding to the learning presentation information presented using the output device; a familiarity calculation unit that performs an analytical process on the product data acquired by the data acquisition unit to calculate the familiarity of the user's knowledge regarding the target content; and a timing setting unit that sets the next presentation timing of the target content in accordance with the familiarity calculated by the familiarity calculation unit.

Images