Information processing method, information processing system, image display method, and image display system

Abstract

To provide a technique for proposing an idea which is different in tendency from existing ideas.SOLUTION: An information processing apparatus is configured to: acquire a plurality of idea vectors obtained by vectorizing a plurality of idea texts based on inputs from a user; execute processing to classifying the idea texts into multiple groups in terms of semantic similarity, based on the idea vectors; determine a specific group that meets a determination condition indicating that the number of classified idea texts is small, based on a result of the processing; and execute presentation processing to present a reference text stored in advance in a memory and classified into the specific group, so as to be highlighted.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present disclosure relates to a technology for supporting the generation of ideas. [Background technology]

[0002] Patent document 1 discloses a technology that extracts a first set of context-forming words, which is a set of words that appear close to each other, from a string of characters proposed during a meeting, searches the Internet for information containing strings that include words that make up the first set of context-forming words, extracts a second set of context-forming words by analyzing information containing the obtained string, and presents a second keyword extracted based on the second set of context-forming words. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 5123591 Summary of the Invention [Problem to be solved by the invention]

[0004] However, in the above conventional technology, since keywords that are strongly related to the topic of the meeting are extracted as the second keywords, there is a problem in that it is not possible to propose keywords that have a different tendency from the ideas already proposed.

[0005] The present disclosure aims to provide a technology that can propose ideas that are different in tendency from previously proposed ideas. [Means for solving the problem]

[0006] An information processing method in one aspect of the present disclosure is an information processing method in a computer, which includes obtaining a plurality of idea vectors by vectorizing a plurality of idea texts based on input from a user, performing a process of grouping the plurality of idea texts into a plurality of groups based on semantic closeness based on the plurality of idea vectors, determining a specific group that satisfies a decision condition indicating that the number of classified idea texts is small based on the result of the grouping process, and performing a presentation process to present reference texts that are pre-stored in a memory and classified into the specific group in an highlighted manner. Effect of the Invention

[0007] This configuration makes it possible to propose ideas that are different in tendency from previously proposed ideas. [Brief description of the drawings]

[0008] [Figure 1] 1 is a diagram illustrating an example of a network configuration of an idea providing system according to an embodiment of the present disclosure. [Diagram 2] FIG. 2 is a block diagram showing an example of a detailed configuration of the idea providing system. [Diagram 3] 3A to 3C are diagrams illustrating an example of data configurations of a dictionary table, an idea sentence table, a group table, and a proposal candidate table. [Figure 4] 13 is a flowchart showing an example of a process for generating a word map in which idea words or idea sentences are highlighted. [Diagram 5] FIG. 1 illustrates an example process for generating a word map with highlighted reference text. [Figure 6] 13 is a flowchart illustrating an example of a process of a proposal switching unit. [Figure 7] 6 is a flowchart showing details of the deletion process shown in step S106 in FIG. 5. [Figure 8] 6 is a flowchart showing details of the addition process shown in step S107 in FIG. 5. [Figure 9] 6 is a flowchart showing details of a specific group determination process shown in step S111 of FIG. 5. [Figure 10] 13 is a frequency distribution graph illustrating a third example of the decision condition. [Figure 11] FIG. 5 is a sequence diagram corresponding to the flowchart shown in FIG. [Figure 12] FIG. 6 is a sequence diagram corresponding to the flowchart shown in FIG. [Figure 13] FIG. 7 is a sequence diagram corresponding to the flowchart shown in FIG. [Figure 14] FIG. 4 is a diagram showing an example of a display image displayed on a monitor. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] (Foundational knowledge of this disclosure) In recent years, research has been conducted into technology that supports idea generation in meetings where participants are stuck on idea generation. In meetings, if participants suggest words that have not been mentioned in the meeting, ideas are semi-forcefully spread, which makes the meeting more lively.

[0010] In the technology of Patent Document 1, as described above, a second keyword is presented. This second keyword is extracted based on a second context-forming word set. However, this second context-forming word set is a set of words obtained by analyzing information including a character string including words constituting a first context-forming word set, which is a set of words that appear close to each other in a character string proposed during a meeting. Therefore, in the technology of Patent Document 1, a keyword that is strongly related to the topic of the meeting is likely to be presented as the second keyword. As a result, the technology of Patent Document 1 has a problem in that it cannot propose ideas that have a different tendency from the keywords already proposed.

[0011] The present disclosure has been made to solve such problems, and aims to provide a technology that can present ideas that are different in tendency from previously presented ideas.

[0012] (1) An information processing method in one aspect of the present disclosure is an information processing method in a computer, including: obtaining a plurality of idea vectors obtained by vectorizing a plurality of idea texts based on an input from a user; executing a process of grouping the plurality of idea texts into a plurality of groups based on semantic closeness based on the plurality of idea vectors; determining a specific group that satisfies a determination condition indicating that the number of classified idea texts is small based on a result of the grouping process; and executing a presentation process to present reference texts that are pre-stored in a memory and classified into the specific group in an highlighted manner.

[0013] According to this configuration, a plurality of idea texts are classified into a plurality of groups based on their similarity in meaning, a specific group that satisfies a determination condition indicating that the number of classified idea texts is small is determined from among the plurality of groups, and reference texts classified into the specific group are presented in an emphasized manner. Therefore, ideas with a different tendency from previously presented ideas can be presented.

[0014] (2) In the information processing method described in (1), the grouping process may include classifying the plurality of idea texts so that they correspond to any one of a plurality of basis vectors based on the plurality of idea vectors.

[0015] In this case, the grouping can be performed accurately.

[0016] (3) In the information processing method according to (1) or (2) above, the number of idea texts included in the specific group is y, a predetermined lower limit is x1, and a predetermined upper limit is x2. The determination condition is <y≦x2であってもよい。

[0017] In this case, a reference text can be presented that is different from previous ideas but does not deviate too far from the topic of the meeting.

[0018] (4) In the information processing method described in any one of (1) to (3) above, the specific group includes a first specific group and a second specific group, the presentation process includes a first presentation process of presenting a first reference text classified into the first specific group in a highlighted manner, and a second presentation process of displaying a second reference text classified into the second specific group in a highlighted manner, and the first presentation process and the second presentation process may be switched in response to a switching operation from the user.

[0019] In this case, the reference text included in either the first specific group or the second specific group can be highlighted in response to a switching operation.

[0020] (5) In the information processing method described in (4) above, the second specific group may be a small group having a smaller number of classified idea texts than the first specific group, or a large group having a larger number of classified idea texts than the first specific group, and in the switching operation, a graphical user interface that allows the user to select one of a first switch and a second switch may be presented to the user, and when the first switch is selected by the user, the first presentation process may be executed, and when the second switch is selected by the user, the second presentation process may be executed.

[0021] The more idea texts a particular group has, the closer the reference texts classified into that particular group are to the meeting agenda, and the fewer idea texts a particular group has, the farther the reference texts classified into that particular group are from the meeting agenda. In this configuration, when the first switching is selected, the first reference text is presented in an emphasized manner, and when the second switching is selected, the second reference text is presented in an emphasized manner. Therefore, in response to the switching operation, the reference text that has an appropriate proximity to the meeting agenda can be displayed in an emphasized manner. In addition, since the switching operation is performed through a graphical user interface, the user can easily input the switching operation.

[0022] (6) In the information processing method according to any one of (1) to (5) above, the input from the user may be a voice of the user.

[0023] In this case, the idea text can be input by voice, which makes it easier to input the idea text.

[0024] (7) In the information processing method described in any one of (1) to (2) and (4) to (6) above, the determination condition may be a condition that the specific group is determined to be a group in which the number of idea texts is equal to a value obtained by multiplying the number of idea texts included in the group with the largest number of idea texts by a coefficient greater than 0 and less than 1.

[0025] In this case, reference text that is different in tendency from previous ideas but does not deviate too far from the topic of the meeting can be presented in an emphasized manner.

[0026] (8) In the information processing method described in any one of (1) to (2) and (4) to (6) above, the determination condition may be a condition that the specific group is determined to be a group located at the lower Z percentile in a frequency distribution graph in which the multiple groups are arranged in descending order of the number of idea texts.

[0027] In this case, reference text that is different in tendency from previous ideas but does not deviate too far from the topic of the meeting can be presented in an emphasized manner.

[0028] (9) In the information processing method according to any one of (1) to (8) above, the grouping process may include calculating a plurality of orthogonal basis vectors from the plurality of idea vectors, the number of which is equal to the number of components of the idea vectors, and classifying each of the plurality of idea vectors into an orthogonal basis vector that has a smallest distance among the plurality of orthogonal basis vectors.

[0029] In this case, the idea text is classified into one of a plurality of orthogonal basis vectors obtained from the idea vector, so that the idea text can be classified more accurately.

[0030] (10) In the information processing method described in (9) above, calculation of the multiple orthogonal basis vectors may include, when the number of the multiple idea texts is insufficient for the number of components of the idea vector, generating a pseudo idea vector having the insufficient number of components using random numbers, the pseudo idea vector being linearly independent from the multiple idea vectors, and calculating the multiple orthogonal basis vectors using the pseudo idea vector and the multiple idea vectors.

[0031] In this case, even if the number of idea vectors is insufficient for the number of components of the idea vector, pseudo idea vectors are generated to make up for the shortage, so that orthogonal basis vectors equal to the number of components of the idea vector can be prepared, thereby enabling more accurate classification of the idea vectors.

[0032] (11) In the information processing method described in (9) above, calculation of the multiple orthogonal basis vectors may include, when the number of the multiple idea texts exceeds the number of components of the idea vector, deleting idea vectors with an excess number of components so as to diversify the distribution of the multiple idea texts.

[0033] In this case, even if the number of idea vectors exceeds the number of components of the idea vector, an orthogonal basis vector for the number of components of the idea vector can be prepared without compromising the distribution characteristics of the idea vector, and the idea vector can be classified more accurately.

[0034] (12) An information processing system in another aspect of the present disclosure is an information processing system including a processor, wherein the processor acquires a plurality of idea vectors obtained by vectorizing a plurality of idea texts based on input from a user, performs a process of grouping the plurality of idea texts into a plurality of groups based on semantic closeness based on the plurality of idea vectors, determines a specific group that satisfies a decision condition indicating that the number of classified idea texts is small based on a result of the grouping process, and performs a presentation process of presenting reference texts that are pre-stored in a memory and classified into the specific group in an highlighted manner.

[0035] According to this configuration, it is possible to provide an information processing system that can present ideas that have a different tendency from previously presented ideas.

[0036] (13) An image display method in another aspect of the present disclosure is an image display method in a computer, comprising: obtaining a plurality of idea vectors obtained by vectorizing a plurality of idea texts based on input from a user; displaying the plurality of idea texts at positions corresponding to each component of the corresponding plurality of idea vectors; generating a display image in which the plurality of reference texts are displayed at positions corresponding to each component of a plurality of reference text vectors obtained by vectorizing a plurality of reference texts pre-stored in a memory; and displaying the display image on a display, wherein the display image highlights sparse areas in which the plurality of idea texts are sparse.

[0037] According to this configuration, in a display image in which a plurality of idea texts and a plurality of reference texts are displayed at the positions of vectorized components, sparse regions in which a plurality of idea texts are sparse are highlighted. Therefore, ideas that have a different tendency from previously presented ideas can be presented.

[0038] (14) In the image display method described in (13) above, the displayed image may include a selection unit that accepts selection of one of a first switch and a second switch, a first sparse region in which the plurality of idea texts is denser than the sparse region, and a second sparse region in which the plurality of idea texts is sparser than the sparse region, and when the first switch is selected, the highlighting is switched to the first sparse region, and when the second switch is selected, the highlighting is switched to the second sparse region.

[0039] In this case, in response to a switching operation, reference texts having appropriate proximity to the meeting topic can be displayed in an emphasized manner.

[0040] (15) In the image display method described in (13) or (14) above, the displayed image may include a reference text display section that displays the plurality of reference texts included in the sparse region in a list format.

[0041] In this case, multiple reference texts contained in the sparse region can be easily identified.

[0042] (16) In the image display method according to any one of (13) to (15) above, the highlighting may be performed by displaying the sparse region in a color different from a color of an area outside the sparse region.

[0043] In this case, the reference text included in the sparse region can be easily distinguished from the reference text not included in the sparse region.

[0044] (17) An image display system in another aspect of the present disclosure is an image display system including a processor, wherein the processor acquires a plurality of idea vectors obtained by vectorizing a plurality of idea texts based on input from a user, displays the plurality of idea texts at positions corresponding to each component of the corresponding plurality of idea vectors, and generates a display image displaying the plurality of reference texts at positions corresponding to each component of a plurality of reference text vectors obtained by vectorizing a plurality of reference texts pre-stored in a memory, displays the display image on a display, and highlights sparse areas where the plurality of idea texts are sparse.

[0045] According to this configuration, it is possible to provide a display device that can present reference texts that have a different tendency from previously presented ideas.

[0046] (18) An information processing method in yet another aspect of the present disclosure includes acquiring voice data output from a microphone that detects voice from a user, converts the detected voice into an electrical signal, and generates voice data from the converted electrical signal; generating a plurality of idea texts based on the voice data; generating a plurality of idea vectors by vectorizing the plurality of idea texts; performing a process of grouping the plurality of idea texts into a plurality of groups based on semantic closeness based on the plurality of idea vectors; determining a specific group that satisfies a determination condition indicating that the number of classified idea texts is small based on a result of the grouping process; and performing a presentation process to present reference texts that are pre-stored in a memory and classified into the specific group in an highlighted manner.

[0047] The present disclosure can also be realized as an information processing program that causes a computer to execute each of the characteristic components included in such an information processing system. Also, the present disclosure can also be realized as an image display program that causes a computer to execute each of the characteristic components included in such an image display system.

[0048] Furthermore, it goes without saying that such computer programs can be distributed on non-transitory computer-readable recording media such as CD-ROMs or via communications networks such as the Internet.

[0049] Note that each of the embodiments described below shows a specific example of the present disclosure. The numerical values, shapes, components, steps, and order of steps shown in the following embodiments are merely examples and are not intended to limit the present disclosure. Furthermore, among the components in the following embodiments, components that are not described in an independent claim showing a top-level concept are described as optional components. Furthermore, in all of the embodiments, the respective contents can be combined.

[0050] (Embodiment) FIG. 1 is a diagram showing an example of a network configuration of an idea providing system according to an embodiment of the present disclosure. The idea providing system is a system that aims to invigorate a meeting by presenting a reference text that has a different tendency from existing ideas. The idea proposal system includes a terminal 100, an idea providing server 200, a storage unit 300, and a monitor 400. The terminal 100 to the monitor 400 are connected to each other via a network so that they can communicate with each other. The network is a wide area communication network including, for example, the Internet and a mobile phone communication network. In FIG. 1, the terminal 100 and the monitor 400 are installed in a space where a meeting is held. An example of the space may be a conference hall where participants gather. In the case of a WEB conference, an example of the space may be a space where multiple participants (users) who participate in the WEB conference are located.

[0051] The terminal 100 is a computer used for inputting an idea sentence. An idea sentence is a sentence that indicates an idea related to the agenda of a meeting and is proposed by a user. Hereinafter, words included in an idea sentence are referred to as idea words. Idea words are an example of idea text. In this embodiment, the idea sentence is input by voice. However, this is only an example, and the idea sentence may be input using an input device such as a keyboard or a touch panel.

[0052] The terminal 100 may be, for example, a portable computer such as a smartphone or a tablet computer, or may be a desktop computer. The terminal 100 includes a microphone, a keyboard, a USB memory, a monitor, and the like.

[0053] The microphone acquires voice data of the idea sentence spoken by the user. In detail, the microphone detects the voice from the user, converts the detected voice into an electrical signal, and generates the voice data from the converted electrical signal.

[0054] The keyboard accepts input of idea sentences by the user. The monitor 400 is composed of a display device such as a liquid crystal display panel, and displays a display image 1500 (see FIG. 15) that supports the creation of ideas. The monitor 400 may be a display of the terminal 100, or may be composed of a computer separate from the terminal 100.

[0055] The idea providing server 200 is configured with a computer such as a cloud server. The storage unit 300 is configured with a computer equipped with a rewritable non-volatile storage device such as a hard disk drive (HDD) and a solid state drive (SSD). The storage unit 300 stores a dictionary table 311, an idea sentence table 312, a group table 313, and a proposal candidate table 314 shown in FIG. 3. The storage unit 300 may be included in the idea providing server 200. The idea providing server 200 displays a display image 1500 on a monitor 400.

[0056] 2 is a block diagram showing an example of a detailed configuration of the idea providing system. The terminal 100 includes an input unit 101 and a communication unit .

[0057] The input unit 101 is composed of a microphone, a keyboard, etc., as shown in Fig. 1. The input unit 101 acquires an idea sentence. The communication unit 102 is composed of a communication circuit that connects the terminal 100 to a network.

[0058] The idea providing server 200 includes a sentence analysis unit 201, a screen generation unit 202, a vector analysis unit 203, an idea proposal unit 204, a proposal switching unit 205, a communication unit 206, and a storage unit 207. The sentence analysis unit 201 to the proposal switching unit 205 are configured, for example, by a central processing unit (CPU). The sentence analysis unit 201 to the proposal switching unit 205 are realized by the CPU executing an information processing program. However, this is just one example, and the sentence analysis unit 201 to the proposal switching unit 205 may be configured by a dedicated hardware circuit. The CPU is an example of a processor.

[0059] The sentence analysis unit 201 acquires the idea sentence transmitted from the terminal 100 using the communication unit 206. The sentence analysis unit 201 stores the acquired idea sentence in the memory unit 301 of the storage unit 300. The sentence analysis unit 201 extracts idea words from the idea sentence by analyzing the idea sentence. The sentence analysis unit 201 converts the idea words into word vectors. Hereinafter, the idea words converted into word vectors are referred to as idea vectors.

[0060] The terminal 100 may transmit voice data generated by a microphone to the idea providing server 200. In this case, the sentence analysis unit 201 may generate an idea sentence by converting the voice data transmitted from the terminal 100 into text data.

[0061] Furthermore, the terminal 100 may convert voice data generated by the microphone into text data and transmit the converted text data to the idea providing server 200. In this case, the sentence analysis unit 201 may obtain the text data transmitted from the terminal 100 as an idea sentence.

[0062] The screen generating unit 202 generates display data for displaying a word map on the monitor 400 in which an area in which a plurality of idea sentences or a plurality of idea words are located is highlighted, and transmits the generated display data to the monitor 400 using the communication unit 206. The screen generating unit 202 acquires from the monitor 400, using the communication unit 206, a selection instruction by the user as to whether to display idea words or idea sentences on the word map.

[0063] The vector analysis unit 203 acquires the idea sentence stored in the storage unit 300 by using the communication unit 206. The vector analysis unit 203 acquires a plurality of idea vectors by analyzing the acquired idea sentence, similar to the sentence analysis unit 201. The vector analysis unit 203 executes a grouping process for grouping a plurality of idea words into a plurality of groups according to their semantic similarity, based on the acquired plurality of idea vectors.

[0064] The grouping process includes classifying the plurality of idea words based on the plurality of idea vectors so that the plurality of idea words correspond to any of the plurality of basis vectors. The basis vectors can be orthogonal basis vectors obtained from the plurality of idea vectors.

[0065] In detail, the grouping process includes calculating a plurality of orthogonal basis vectors from a plurality of idea vectors, the number of which is equal to the number of components of the idea vectors, and classifying each of the plurality of idea vectors into an orthogonal basis vector that has the smallest distance among the plurality of orthogonal basis vectors. The number of components of an idea vector is the number of components (number of dimensions) of the idea vector. For example, if the idea vector is a 300-dimensional vector, 300 orthogonal basis vectors are generated. The process of calculating the orthogonal basis vectors from a plurality of idea vectors can employ, for example, the Gram-Schmidt orthogonalization method. Any distance that evaluates the closeness of vectors may be employed as the distance, and for example, cosine distance, Euclidean distance, etc. can be employed.

[0066] When the number of the multiple idea words is insufficient for the number of components of the idea vector, the vector analysis unit 203 generates a pseudo idea vector with the missing number of components and linearly independent from the multiple idea vectors using random numbers. Then, the vector analysis unit 203 calculates multiple orthogonal basis vectors using the pseudo idea vector and the multiple idea vectors. For example, when the number of components of the idea vector is 300 and the number of idea words is 290, the missing number of components is 10. In this case, 10 pseudo idea vectors are generated. A linearly independent pseudo idea vector refers to a vector that is not an integer multiple of any of the existing multiple idea vectors.

[0067] When the number of multiple idea words exceeds the number of components of an idea vector, the vector analysis unit 203 deletes the idea vector with the excess number of components so that the distribution of multiple idea words is dispersed. For example, when the number of components of an idea vector is 300 and the number of idea words is 310, the excess number of components is 10. In this case, 10 idea vectors are deleted.

[0068] The Idea Proposal Department 204 determines a specific group that satisfies a determination condition indicating that the number of classified idea words is small based on the result of the grouping process.

[0069] In the first example of the determination condition, assuming the number of idea words included in a specific group is y, a predetermined lower limit value is x1, and a predetermined upper limit value is x2, then x1 < y ≤ x2. In this case, the specific group is determined from among the groups where the number of idea words is greater than x1 and less than or equal to x2. Note that as the values of the lower limit x1 and the upper limit x2, values preset by the user can be adopted. The first example of the determination condition may be composed of not a numerical range but a single value. As the lower limit value x1 and the upper limit value x2, appropriate values such as x1 = 1, x2 = 2, or x1 = 2, x2 = 4 can be adopted. Also, the lower limit value x1 and the upper limit value x2 may be represented by a function such as 1 ≤ x1 ≤ 3 and x2 = x1 + z (1 ≤ z ≤ 5, z is an integer).

[0070] The second example of the determination condition may be a condition that a group with the same number of idea words as the value obtained by multiplying the number of idea words in the group with the largest number of idea words by a coefficient greater than 0 and less than 1 is determined as the specific group. For example, assuming the number of idea words in the largest group is ymax, the second example of the determination condition is represented by ymax·α. However, α is a real number greater than 0 and less than 1. In this case, the specific group is determined from among the groups where the number of idea words is ymax·α.

[0071] The third example of the determination condition is a condition that a group located at the lower Z% point in the frequency distribution of idea words for each of a plurality of groups is determined as the specific group. FIG. 10 is a frequency distribution graph for explaining the third example of the determination condition. In this frequency distribution graph, the vertical axis represents the frequency and the horizontal axis represents the group. Note that the groups are arranged on the horizontal axis in descending order of the number of idea words. P1 indicates the lower Z percent point. The lower Z percent point P1 indicates a point on the horizontal axis where the ratio of the total number B of the lower frequencies to the total number A of the frequencies is Z percent. An example of Z is 20, but appropriate values such as 30 and 10 can be adopted.

[0072] Returning to FIG. 2, the idea suggestion unit 204 executes a presentation process to present, in an emphasized manner, reference texts that are pre-stored in the storage unit 301 (an example of a memory) of the storage unit 300 and that are classified into a specific group. The reference texts refer to words pre-stored in the dictionary table 311 shown in FIG. 3. The idea suggestion unit 204 plots the reference texts on a word map, generates display data for a display image including a word map that highlights reference texts included in a specific group among the plotted reference texts, and transmits the generated display data to the monitor 400 using the communication unit 206. As a result, a display image 1500 as shown in FIG. 15 is displayed on the monitor 400.

[0073] Returning to Fig. 2, the specific group may include a first specific group and a second specific group. In this case, the suggestion switching unit 205 may switch between a first presentation process in which a first reference text classified into the first specific group is presented in an emphasized manner and a second presentation process in which a second reference text classified into the second specific group is presented in an emphasized manner in response to a switching operation by the user. This switching operation is performed using the input unit 401 of the monitor 400.

[0074] The proposal switching unit 205 displays a graphical user interface for selecting one of the first switching and the second switching on the display unit 403 of the monitor 400, and causes the user to input a switching operation. When the first switching is selected by the user using the graphical user interface, the proposal switching unit 205 executes a first presentation process, and when the second switching is selected by the user using the graphical user interface, the proposal switching unit 205 executes a second presentation process.

[0075] For example, the second specific group is a small group having a smaller number of classified idea words than the first specific group, or a large group having a larger number of classified idea words than the first specific group.

[0076] In the following description, the first specific group is a specific group having a larger number of classified idea words than the third specific group, which is a specific group in which the reference text is currently displayed in a highlighted manner, and the second specific group is a specific group having a smaller number of classified idea words than the third specific group.

[0077] The more idea words a specific group has, the closer the reference text classified into that specific group is to the meeting agenda, and the less idea words a specific group has, the farther the reference text classified into that specific group is from the meeting agenda. A user who feels that the reference text classified into the third specific group is closer to the meeting agenda selects the second switching. In this case, the reference text classified into the second specific group is highlighted, so the user can move the highlighted reference text farther away from the meeting agenda. On the other hand, a user who feels that the reference text classified into the third group is farther away from the meeting agenda selects the first switching. In this case, the reference text classified into the first specific group is highlighted, so the user can move the highlighted reference text closer to the agenda.

[0078] The communication unit 206 is composed of a communication circuit that connects the idea providing server 200 to a network. The communication unit 206 transmits display data for displaying a word map to the monitor 400. The communication unit 206 transmits display data for displaying a display image to the monitor 400. The communication unit 206 receives voice data or text data of a text sentence from the terminal 100.

[0079] The storage unit 207 is configured with a rewritable non-volatile storage device such as a solid state drive (SSD) or a hard disk drive (HDD), and stores an information processing program that causes a computer to function as the idea providing server 200.

[0080] The storage unit 300 includes a storage unit 301 and a communication unit 302. The storage unit 301 is configured with a rewritable non-volatile storage device such as a solid state drive (SSD) or a hard disk drive (HDD). The storage unit 301 stores a dictionary table 311, an idea sentence table 312, a group table 313, and a proposal candidate table. FIG. 3 is a diagram showing an example of the data configuration of the dictionary table 311, the idea sentence table 312, the group table 313, and the proposal candidate table 314.

[0081] The dictionary table 311 stores reference texts and reference text vectors obtained by vectorizing the reference texts in association with each other. For example, a large number of words registered in a dictionary or the like are used as the reference texts. In this example, words such as "simmer" and "boil" are registered in the dictionary table 311. For example, word vectors such as word2vec can be used as the reference text vectors. The reference text vectors are generated by inputting the reference texts into a trained model that has been machine-learned in advance to vectorize words. In this example, the reference text vectors are expressed by a component number of 300 dimensions, but this is just an example and may be expressed by a component number other than 300 dimensions.

[0082] The idea sentence table 312 stores idea sentences input to the terminal 100 in association with the date on which the idea sentence was input. In this example, "Boil the vegetables" was uttered at 11:01, so the text data of "Boil the vegetables" is stored in association with the utterance time of 11:01. The idea sentence table 312 may be generated for each conference. In this case, the idea sentence table 312 is associated with the conference ID.

[0083] The group table 313 is a table that stores idea words in association with which group they belong to. The group table 313 stores idea words in association with the group IDs of the groups to which the idea words belong. For example, the idea word "vegetables" belongs to a group with a group ID "e1", so it is stored in association with "e1". For example, an orthogonal basis vector corresponding to a group can be used as the group ID. The group table 313 may be generated for each meeting. In this case, the group table 313 is associated with a meeting ID.

[0084] The proposal candidate table 314 stores a group to which a reference text belongs and a ranking of the reference text in the group in association with each other. Specifically, the proposal candidate table 314 stores a group ID indicating a group to which a reference text belongs, a ranking of the reference text in the group, and the reference text in association with each other. For example, the record in the first row of the proposal candidate table 314 records that the reference text "chest of drawers" belongs to a group with a group ID "e1" and is ranked "1st" in the group. The smaller the distance between the reference text and the orthogonal basis vector that defines the group to which it belongs, the higher the ranking of the reference text is assigned.

[0085] Returning to Fig. 2, the communication section 302 is composed of a communication circuit that connects the storage unit 300 to a network.

[0086] The monitor 400 includes an input unit 401, a screen control unit 402, and a display unit 403. The input unit 401 is composed of input devices such as a touch panel, a keyboard, and a mouse. The input unit 401 accepts the above-mentioned first switching and second switching, etc. The screen control unit 402 acquires display data transmitted from the idea providing server 200, generates a display image 1500 using the acquired display data, and displays the display image 1500 on the display unit 403. The display unit 403 is composed of a display device such as a liquid crystal display.

[0087] The above is the configuration of the idea providing system. Next, the processing of the idea providing system will be described. Fig. 4 is a flow chart showing an example of processing for generating a word map in which idea words or idea sentences are highlighted.

[0088] First, in step S1, the sentence analysis unit 201 acquires an idea sentence input to the terminal 100 from the terminal 100 using the communication unit 206, and transmits the acquired idea sentence to the storage unit 300 using the communication unit 206, thereby storing the idea sentence in the storage unit 300. Here, since the voice data of the idea sentence is transmitted from the terminal 100, the sentence analysis unit 201 converts the voice data of the idea sentence into text data, and transmits the converted text data of the idea sentence to the storage unit 300. As a result, the idea sentence is stored in the idea sentence table 312 shown in Fig. 3. When a conference ID is assigned to the idea sentence table 312, the sentence analysis unit 201 stores the idea sentence in the idea sentence table 312 corresponding to the conference ID acquired in advance.

[0089] Next, in step S2, the sentence analysis unit 201 acquires the idea sentence from the storage unit 300. For example, the sentence analysis unit 201 may acquire the idea sentence by transmitting an acquisition request to the storage unit 300 to request acquisition of the idea sentence. For example, the sentence analysis unit 201 may acquire the idea sentence from the storage unit 300 at a predetermined time interval. In this case, the sentence analysis unit 201 may identify an unacquired idea sentence from the time included in the idea sentence table 312, and acquire the identified unacquired idea sentence from the storage unit 300. When a conference ID is assigned to the idea sentence table 312, the sentence analysis unit 201 may acquire the idea sentence from the idea sentence table 312 corresponding to the conference ID acquired in advance.

[0090] Next, in step S3, the sentence analysis unit 201 obtains the dictionary table 311 from the storage unit 300. This dictionary table 311 is used when determining a reference text in the flowchart of FIG.

[0091] Next, in step S4, the sentence analysis unit 201 performs morphological analysis on the idea sentence acquired in step S2 to separate the acquired idea sentence into words. Word separation is a process of dividing a sentence into words. For example, a text sentence "Play straddling the pattern part" is separated into words as "play straddling the pattern part".

[0092] Next, in step S5, the sentence analysis unit 201 extracts the part of speech of each word separated from the idea sentence. The part of speech can be specified by morphological analysis.

[0093] Next, in step S6, the sentence analysis unit 201 excludes excluded words from the idea sentence. The excluded words are words having parts of speech registered in the part-of-speech exclusion list. The part-of-speech exclusion list stores in advance excluded parts of speech (e.g., particles) that become noise when evaluating an idea sentence. By excluding the excluded words from the idea sentence, evaluation words remain. The evaluation words are words consisting of parts of speech such as nouns, adjectival verbs, verbs, and adjectives. These remaining evaluation words become idea words.

[0094] In step S7, the sentence analysis unit 201 inputs the idea words into a trained model for vectorizing the idea words, and converts the idea words into idea vectors.

[0095] In step S8, the screen generating unit 202 accepts a selection from the user as to whether to display idea sentences or idea words on the word map. In detail, the screen generating unit 202 displays a graphical user interface for making this selection on the monitor 400, and receives a selection instruction inputted using this graphical user interface from the monitor 400, thereby accepting the above selection.

[0096] If the selection instruction indicates the selection of an idea sentence (YES in step S8), the process proceeds to step S10, and if the selection instruction indicates the selection of an idea word (NO in step S8), the process proceeds to step S9.

[0097] Next, in step S9, the screen generating unit 202 displays the word map in which the idea words are highlighted on the monitor 400. In detail, the screen generating unit 202 compresses the idea vectors corresponding to the idea words into a number of dimensions that can be visualized (for example, two dimensions), and displays the idea words at the positions on the word map indicated by the dimensionally compressed idea vectors. The screen generating unit 202 also displays the reference texts at the positions on the word map similar to the idea words. Then, the screen generating unit 202 highlights the idea words among the idea words and the reference texts displayed on the word map. As a mode of highlighting, a mode of displaying the area where the idea words are located in a heat map shape in a color different from the surrounding area can be adopted. As a mode of dimensional compression, Principal Component Analysis (PCA), tSNE (t-Distributed Stochastic Neighbor Embedding), UMAP, etc. can be adopted.

[0098] Next, in step S10, the screen generating unit 202 generates a sentence vector by vectorizing the idea sentence. The sentence vector is generated by combining a plurality of idea vectors indicating a plurality of idea words that constitute the idea vector.

[0099] Next, in step S11, the screen generating unit 202 displays the word map with the idea sentence highlighted on the monitor 400. In detail, the screen generating unit 202 dimensionally compresses the sentence vector to a visualizable number of components, and displays the idea sentence at a position on the word map indicated by the dimensionally compressed sentence vector. The screen generating unit 202 also displays the reference text at a position on the word map similar to the idea word. The screen generating unit 202 then highlights the idea sentence among the idea sentence and the reference text displayed on the word map.

[0100] 5 is a diagram showing an example of a process for generating a word map in which reference text is highlighted. This flowchart is executed after the process in FIG. 4 is completed. First, in step S101, the vector analysis unit 203 acquires an idea sentence from the storage unit 300. The details of this process are the same as those in step S2 in FIG. 4.

[0101] Next, in step S102, the vector analysis unit 203 extracts idea words from the idea sentence. Details of this process are the same as steps S4 to S6 in FIG.

[0102] Next, in step S103, the vector analysis unit 203 converts the idea words into idea vectors. The details of this process are the same as those in step S7 of FIG.

[0103] Next, in step S104, the vector analysis unit 203 judges whether the number of idea words is equal to the number of components of the idea vector. If the number of idea words is equal to the number of components of the idea vector (YES in step S104), the process proceeds to step S108. On the other hand, if the number of idea words is not equal to the number of components of the idea vector (NO in step S104), the process proceeds to step S105.

[0104] Next, in step S105, the vector analysis unit 203 judges whether the number of idea words is greater than the number of components of the idea vector. If the number of idea words is greater than the number of components of the idea vector (YES in step S105), the vector analysis unit 203 executes a deletion process (step S106). The details of the deletion process will be described later.

[0105] On the other hand, if the number of idea words is equal to or less than the number of components of the idea vector (NO in step S105), the vector analysis unit 203 executes an addition process (step S107). The details of the addition process will be described later.

[0106] Next, in step S108, the vector analysis unit 203 uses the Gram-Schmidt orthogonalization method to determine, from the idea vectors of the number of components of the idea vector, orthogonal basis vectors of the number of components.

[0107] Next, in step S109, the vector analysis unit 203 executes a grouping process to classify the multiple idea words into one of multiple groups defined by the orthogonal basis vectors. In detail, the vector analysis unit 203 determines an orthogonal basis vector with a minimum distance among the multiple orthogonal basis vectors for each of the multiple idea words, and classifies the multiple idea words into the determined orthogonal basis vector. Then, the vector analysis unit 203 stores the classification result in the group table 313. If the group table 313 is associated with a conference ID, the vector analysis unit 203 may store the classification result in the corresponding group table 313.

[0108] Next, in step S110, the idea proposing unit 204 classifies the multiple reference texts stored in the dictionary table 311 into the groups generated in step S109, and stores the classification results in the proposal candidate table 314. In detail, for each of the multiple reference texts stored in the dictionary table 311, the idea proposing unit 204 identifies an orthogonal basis vector with a minimum distance from among the multiple orthogonal basis vectors determined in step S108, and classifies the reference text into a group defined by the identified orthogonal basis vector. Note that, if the proposal candidate table 314 is associated with a conference ID, the idea proposing unit 204 may store the classification results in the corresponding proposal candidate table 314.

[0109] Next, in step S111, the idea suggesting unit 204 executes a process for determining a specific group. The details of this process will be described later.

[0110] Next, in step S112, the idea suggestion unit 204 generates a word map that plots the reference text at a position on the word map specified by the reference text vector and highlights the reference text included in a specific group among the plotted reference texts. Then, the idea suggestion unit 204 displays a display image 1500 including the generated word map on the monitor 400. At this time, the idea suggestion unit 204 also plots the idea words at a position on the word map specified by the idea vector.

[0111] FIG. 6 is a flowchart showing an example of the processing of the proposal switching unit 205. First, in step S201, the proposal switching unit 205 accepts the selection of a "close" button or a "far" button included in the display image. As shown in FIG. 15, a "close" button 1541 and a "far" button 1542 are graphical user interfaces provided in the display image 1500. The "close" button 1541 is a button selected when the user feels that the currently highlighted reference text is close to the agenda of the meeting. This selection corresponds to the second switching described above. The "far" button 1542 is a button selected when the user feels that the currently highlighted reference text is far from the agenda of the meeting. This selection corresponds to the first switching described above.

[0112] Next, in step S202, the proposal switching unit 205 acquires the group table 313 from the storage unit 300, and acquires the number of idea words included in each group by referring to the acquired group table 313.

[0113] Next, in step S203, the proposal switching unit 205 determines whether the "near" button 1541 or the "far" button 1542 has been selected.

[0114] When the "close" button 1541 is selected (YES in step S203), the suggestion switching unit 205 identifies the second specific group (step S204). In detail, the suggestion switching unit 205 identifies, as the second specific group, each of one or more groups that includes idea words having a number of idea words smaller than the number of idea words m included in the third specific group, which is the specific group in which the reference text is currently highlighted, and includes idea words having a number of words closest to the number of words m. For example, when the number of words m is m=7, if there are two groups each having 6 words, each of these two groups becomes the second specific group.

[0115] On the other hand, when the "Far" button 1542 is selected (NO in step S203), the suggestion switching unit 205 identifies the first specific group (step S205). In detail, the suggestion switching unit 205 identifies each of one or more groups that have a number of idea words larger than the number of words m of idea words included in the third specific group and that have a number of words closest to the number of words m as the first specific group. For example, when the number of words m of the third specific group is m=7, and there are two groups each having 8 words, these two groups become the first specific group.

[0116] Next, in step S206, the proposal switching unit 205 randomly selects one specific group from the first specific group identified in step S204 or the second specific group identified in step S205. This one specific group becomes a new third specific group. Note that, if the number of first specific groups identified in step S204 or second specific groups identified in step S205 is one, the proposal switching unit 205 may select one first specific group or one second specific group as one specific group without executing the process of step S206.

[0117] Next, in step S207, the word map of the display image 1500 is updated so that the reference texts belonging to the one specific group selected in step S206 are highlighted. This switches the highlighted reference texts to reference texts closer or further away from the meeting topic.

[0118] Fig. 7 is a flowchart showing details of the deletion process shown in step S106 in Fig. 5. First, in step S1061, the vector analysis unit 203 calculates the number of excess words n. For example, when the number of dimensions of the idea vector is 300, and the number of idea words is 310, n=10.

[0119] Next, in step S1062, the vector analysis unit 203 generates a dendrogram of all idea words by applying hierarchical clustering to all idea words. The dendrogram is a diagram that hierarchically represents the proximity of idea words to each other.

[0120] Next, in step S1063, the vector analysis unit 203 starts a deletion loop of idea words.

[0121] Next, in step S1064, the vector analysis unit 203 identifies a pair of branches located at the end of the dendrogram, and deletes idea words located at one of the identified pair of branches. Here, the pair of branches located at the end refers to a branch in which a pair of idea words with the smallest similarity is located in the dendrogram. This thins out dense branches in the dendrogram.

[0122] Next, in step S1065, the vector analysis unit 203 checks whether the number of deleted idea words has reached the number of words n. If the number of deleted idea words has not reached the number of words n (NO in step S1065), the process returns to step S1064. On the other hand, if the number of deleted idea words has reached the number of words n (YES in step S1065), the vector analysis unit 203 ends the deletion loop (step S1066) and ends the deletion process.

[0123] In addition, the vector analysis unit 203 may delete the n idea words using a method other than hierarchical clustering. For example, the vector analysis unit 203 may delete the n idea words by searching for a deletion pattern with the maximum variance among all deletion patterns in which the n idea words are deleted from all idea words.

[0124] Fig. 8 is a flowchart showing details of the addition process shown in step S107 in Fig. 5. First, in step S1071, the vector analysis unit 203 calculates the number of words n that are missing in the idea vector. For example, when the number of dimensions of the idea vector is 300 and the number of idea words is 290, n=10.

[0125] Next, in step S1072, the vector analysis unit 203 starts an idea word addition loop.

[0126] Next, in step S1073, the vector analysis unit 203 generates a pseudo idea vector using random numbers. The pseudo idea vector has the same number of components as the idea vector.

[0127] Next, in step S1074, the vector analysis unit 203 determines whether the pseudo idea vector generated in step S1073 is linearly independent of all vectors included in the idea vector list. The idea vector list is a list that records all idea vectors and the pseudo idea vectors added in step S1075, which will be described later.

[0128] If the pseudo idea vector is linearly independent of all vectors recorded in the idea vector list (YES in step S1074), the vector analysis unit 203 adds the pseudo idea vector to the idea vector list. On the other hand, if the pseudo idea vector is not linearly independent of any vector recorded in the idea vector list (NO in step S1074), the vector analysis unit 203 returns the process to step S1073 and generates a pseudo idea vector.

[0129] Next, in step S1076, the vector analysis unit 203 checks whether the number of added pseudo idea vectors has reached the missing number of words n. If the number of added pseudo idea vectors has not reached the missing number of words n (NO in step S1076), the process returns to step S1073. On the other hand, if the number of added pseudo idea vectors has reached the missing number of words n (YES in step S1076), the vector analysis unit 203 ends the addition loop (step S1077) and ends the addition process.

[0130] Fig. 9 is a flowchart showing details of the specific group determination process shown in step S111 of Fig. 5. First, in step S1111, the idea suggesting unit 204 identifies a group that satisfies a determination condition. As the determination condition, any one of the first to third examples described above is adopted.

[0131] When the first example of the determination condition is adopted, the lower limit x1 and the upper limit x2 preset by the user are determined as the determination condition. In this case, a group whose number y of idea words is in the range of greater than the lower limit x1 and less than or equal to the upper limit x2 is identified as a group that satisfies the determination condition.

[0132] If the second example of the decision criterion is adopted, then the group whose number of idea words is equal to the value given by ymax·α above is identified as the group that satisfies the decision criterion.

[0133] When the third example of the decision condition is adopted, the group located at the lower Z percentile P1 described above is identified as the group that satisfies the decision condition.

[0134] Next, in step S1112, the idea suggestion unit 204 selects one group from among the groups that satisfy the determination condition as the specific group. For example, when multiple groups satisfy the determination condition, the idea suggestion unit 204 may select one group randomly determined from the multiple groups as the specific group. Note that when there is only one group that satisfies the specific condition, the one group is determined as the specific group.

[0135] Fig. 11 is a sequence diagram corresponding to the flowchart shown in Fig. 4. In Fig. 11, the same processes as those in Fig. 4 are given the same step numbers. In step S1, an idea sentence is transmitted from the terminal 100 to the storage unit 300 via the idea providing server 200. As a result, the idea sentence is stored in the idea sentence table 312.

[0136] In step S2, the idea sentence is transmitted from the storage unit 300 to the idea providing server 200. In step S3, the dictionary table 311 is transmitted from the storage unit 300 to the idea providing server 200. After that, the processes of steps S4 to S7 are executed by the idea providing server 200, and idea words and idea vectors are obtained from the idea sentence.

[0137] In step S8, the monitor 400 transmits the selection result of whether to display idea sentences or idea words on the word map to the idea providing server 200. In step S9, display data for displaying the word map with idea words highlighted on the monitor 400 is transmitted from the idea providing server 200 to the monitor 400. In step S10, the sentence vector of the idea sentence is generated. In step S11, display data for displaying the word map with idea sentences highlighted on the monitor 400 is transmitted from the idea providing server 200 to the monitor 400. The process of step S9 and the processes of steps S10 and S11 are alternatively performed.

[0138] Fig. 12 is a sequence diagram corresponding to the flowchart shown in Fig. 5. In Fig. 12, the same processes as those in Fig. 5 are assigned the same step numbers. In step S101, an idea statement is transmitted from the storage unit 300 to the idea providing server 200. Thereafter, the processes of steps S102 to S111 are executed by the idea providing server 200, and a specific group is determined. In step S112, display data for displaying the display image 1500 on the monitor 400 is transmitted to the monitor 400.

[0139] Fig. 13 is a sequence diagram corresponding to the flowchart shown in Fig. 6. In Fig. 13, the same processes as those in Fig. 6 are denoted by the same step numbers.

[0140] In step S201, an instruction to select the “close” button 1541 or the “far” button 1542 is transmitted from the monitor 400 to the idea providing server 200. In step S202, the group table 313 is transmitted from the storage unit 300 to the idea providing server 200.

[0141] Thereafter, the processes of steps S203 to S206 are executed by the idea providing server 200, and a group having an appropriate proximity to the agenda of the meeting is selected as a specific group. In step S207, the word map is updated so that the reference text included in the selected specific group is highlighted.

[0142] FIG. 14 is a diagram showing an example of a display image 1500 displayed on the monitor 400. The display image 1500 includes a word map 1510, a group display field 1520, a reference text display field 1530, and a selection section 1540. The word map 1510 displays an idea word or an idea sentence at a position corresponding to a component of a corresponding idea vector. For example, when an idea word is selected in step S8 of FIG. 4, the idea word is displayed in the word map 1510, and when an idea sentence is selected in step S8, the idea sentence is displayed in the word map 1510. The word map 1510 displays a reference text at a position corresponding to a component of a corresponding idea vector. The word map 1510 displays an idea word (or idea sentence) at a position 1511 where a dark dot is plotted, and displays a reference text at a position 1512 where a light dot is plotted.

[0143] Position 1511 is determined by components of an idea vector (or sentence vector) that has been dimensionally compressed to a visualizeable number of dimensions. Position 1512 is determined by components of a reference text vector that has been dimensionally compressed to a visualizeable number of dimensions. For example, the dark dots shown at position 1511 and the idea words (or idea sentences) associated with position 1511 are displayed in a first color (e.g., red), and the light dots shown at position 1512 and the reference text associated with position 1512 are displayed in a second color (e.g., black) different from the first color. Note that the dark dots plotted at position 1511 and the light dots plotted at position 1512 may be omitted.

[0144] The word map 1510 highlights an area 1513 in which a reference text belonging to a specific group is located. In this example, the area 1513 is highlighted in a color (e.g., semi-transparent red) different from the surrounding color. The word map 1510 displays the number of reference texts equal to an upper limit value determined for each group. For example, if the upper limit value is 5, the word map 1510 displays the top 5 reference texts from each group. The top 5 reference texts refer to the reference texts ranked 1st to 5th in the group stored in the proposal candidate table 314. The size of the area 1513 is set to a size that can include all of the positions 1512 of the top 5 reference texts included in the specific group. The area 1513 is an example of a sparse area in which multiple idea texts are sparsely distributed.

[0145] The group display field 1520 displays the group number of the specific group whose reference text is currently highlighted. The group number is a number assigned in advance to identify each group, and is linked to the group ID (orthogonal basis vector). In this example, the group with group number "6" has been determined as the specific group, so the group display field 1520 displays the group number "6."

[0146] The reference text display field 1530 displays in a list format the reference words that belong to a specific group and are displayed in the word map. Here, the top five reference texts that belong to group number "6" are displayed in a list.

[0147] Selection section 1540 includes a "close" button 1541 and a "far" button 1542. The "close" button 1541 is selected when the user feels that the reference text included in area 1513 is close to the agenda of the meeting. The "far" button 1542 is selected when the user feels that the reference text included in area 1513 is far from the agenda of the meeting.

[0148] When the "close" button 1541 is selected, one specific group is identified in the word map 1510 from the above-mentioned second specific group. As a result, the display position of the area 1513 is switched to the display position of the reference text belonging to the identified one specific group, and the reference text that is farther from the meeting agenda is highlighted. The area 1513 that is highlighted after the "close" button 1541 is selected is an example of a "second sparse area in which multiple idea texts are sparser than the sparse area." The selection of the "close" button 1541 is an example of a second switch.

[0149] When the "Far" button 1542 is selected, one specific group is identified in the word map 1510 from the above-mentioned first specific group. As a result, the display position of the area 1513 is switched to the display position of the reference text belonging to the identified one specific group, and the reference text closer to the meeting agenda is highlighted. The area 1513 that is highlighted after the "Far" button 1542 is selected is an example of a "first sparse area in which multiple idea texts are denser than the sparse area." The selection of the "Far" button 1542 is an example of a first switch.

[0150] In this way, according to the present embodiment, a plurality of idea words are classified into a plurality of groups according to their similarity in meaning, a specific group that satisfies a determination condition indicating that the number of classified idea words is small is determined from among the plurality of groups, and reference texts classified into the specific group are presented in an emphasized form. Therefore, ideas that are different in tendency from the ideas already presented can be presented, and the meeting can be revitalized.

[0151] The present disclosure can employ the following modifications.

[0152] (1) In the above embodiment, the sentence analysis unit 201 and the vector analysis unit 203 obtain idea vectors by vectorizing idea words, but the present disclosure is not limited to this, and the idea vectors that have been vectorized in advance may be obtained from, for example, an external device. In this case, the idea providing server 200 does not need to generate idea vectors.

[0153] (2) In step S203 of Fig. 6, when the "close" button 1541 is selected, the second specific group is identified (step S204), and when the "far" button 1542 is selected, the first specific group is identified (step S205), but this is just an example, and when the "close" button 1541 is selected, the first specific group may be identified, and when the "far" button 1542 is selected, the second specific group may be identified. In this case, when the user feels that the currently highlighted reference text is close to the topic of the meeting, the "far" button 1541 is selected, and when the user feels that the currently highlighted reference text is far from the topic of the meeting, the "close" button 1542 is selected.

[0154] (3) The sentence analysis unit 201 to the proposal switching unit 205 are included in the idea providing server 200, but may be distributed and arranged in the terminal 100 and the monitor 400.

[0155] (4) The information processing system of the present disclosure may be configured with at least the idea providing server 200 among the terminal 100, the idea providing server 200, the storage unit 300, and the monitor 400. The image display system of the present disclosure may be configured with at least the idea providing server 200 among the terminal 100, the idea providing server 200, the storage unit 300, and the monitor 400. [Industrial Applicability]

[0156] INDUSTRIAL APPLICABILITY The present disclosure is useful in the technical field of supporting meetings since it can activate meetings. [Explanation of symbols]

[0157] 100: Terminal 200: Idea providing server 201: Sentence analysis department 202: Screen generation section 203: Vector Analysis Section 204: Idea Proposal Department 205: Proposal switching section 206: Communications Department 207: Storage section 300: Storage unit 311: Dictionary table 312: Idea sentence table 313: Group table 314: Proposal candidate table 1500: Display image 1510: Word Map 1530: Reference text display field 1540: Selection section

Claims

1. An information processing method in a computer, comprising: Obtaining a plurality of idea vectors by vectorizing a plurality of idea texts based on an input from a user; Executing a process of grouping the plurality of idea texts into a plurality of groups based on the plurality of idea vectors according to their semantic similarity; determining a specific group that satisfies a determination condition indicating that the number of classified idea texts is small based on the result of the grouping process; and executing a presentation process for presenting, in an emphasized manner, reference texts that are stored in advance in a memory and that are classified into the specific group. Information processing methods.

2. The grouping process includes classifying the plurality of idea texts so as to correspond to any one of a plurality of basis vectors based on the plurality of idea vectors.

2. The information processing method according to claim 1.

3. Let y be the number of idea texts included in the specific group, x1 be a predetermined lower limit, and x2 be a predetermined upper limit. The determination condition is x1<y≦x2.

3. The information processing method according to claim 1 or 2.

4. the specific group includes a first specific group and a second specific group, the presentation process includes a first presentation process of presenting a first reference text classified into the first specific group in a highlighted manner, and a second presentation process of displaying a second reference text classified into the second specific group in a highlighted manner, The first presentation process and the second presentation process are switched in response to a switching operation by the user.

3. The information processing method according to claim 1 or 2.

5. The second specific group is a small group having a smaller number of the classified idea texts than the first specific group, or a large group having a larger number of the classified idea texts than the first specific group, presenting a graphical user interface to the user for selecting one of a first switching and a second switching in the switching operation; When the first switching is selected by the user, the first presentation process is executed; When the second switching is selected by the user, the second presentation process is executed.

5. The information processing method according to claim 4.

6. the input from the user being the user's voice; 3. The information processing method according to claim 1 or 2.

7. The determination condition is a condition that a group having a number of idea texts equal to a value obtained by multiplying the number of idea texts included in the group having the largest number of idea texts by a coefficient greater than 0 and less than 1 is determined as the specific group.

3. The information processing method according to claim 1 or 2.

8. The determination condition is a condition that a group located at a lower Z percentile in a frequency distribution graph in which the plurality of groups are arranged in descending order of the number of idea texts is determined as the specific group.

3. The information processing method according to claim 1 or 2.

9. The grouping process includes: calculating a plurality of orthogonal basis vectors from the plurality of idea vectors, the number of which corresponds to the number of components of the idea vectors; classifying each of the plurality of idea vectors into an orthogonal basis vector having a minimum distance among the plurality of orthogonal basis vectors; 3. The information processing method according to claim 1 or 2.

10. The calculation of the plurality of orthogonal basis vectors includes the steps of: When the number of the plurality of idea texts is insufficient for the number of components of the idea vector, a pseudo idea vector having the insufficient number of components is generated by random numbers, the pseudo idea vector being linearly independent from the plurality of idea vectors; calculating the plurality of orthogonal basis vectors using the pseudo idea vector and the plurality of idea vectors; 10. The information processing method according to claim 9.

11. The calculation of the plurality of orthogonal basis vectors includes, when the number of the plurality of idea texts exceeds the number of components of the idea vector, deleting the idea vector having the excess number of components so that the distribution of the plurality of idea texts is dispersed.

10. The information processing method according to claim 9.

12. An information processing system including a processor, The processor, Obtaining a plurality of idea vectors by vectorizing a plurality of idea texts based on input from a user; Executing a process of grouping the plurality of idea texts into a plurality of groups based on the plurality of idea vectors, according to the similarity of meaning; Based on the result of the grouping process, a specific group is determined that satisfies a determination condition indicating that the number of classified idea texts is small; executing a presentation process for presenting, in an emphasized manner, reference texts that are stored in advance in a memory and that are classified into the specific group; Information processing system.

13. A method for displaying an image in a computer, comprising the steps of: Obtaining a plurality of idea vectors by vectorizing a plurality of idea texts based on an input from a user; Displaying the plurality of idea texts at positions according to the components of the corresponding plurality of idea vectors; and generating a display image in which a plurality of reference texts are displayed at positions corresponding to respective components of a plurality of reference text vectors obtained by vectorizing a plurality of reference texts stored in advance in a memory; displaying the display image on a display; the display image highlights a sparse region where the plurality of idea texts are sparse; Image display method.

14. The display image is A selection unit that accepts selection of one of the first switching and the second switching; a first sparse region in which the plurality of idea texts are denser than the sparse region; a second sparse region in which the plurality of idea texts is sparser than the sparse region; When the first switching is selected, the highlighting is switched to the first sparse region; When the second switching is selected, the highlighting is switched to the second sparse region. The image display method according to claim 13.

15. the display image includes a reference text display field that displays the plurality of reference texts included in the sparse region in a list format; 15. The image display method according to claim 13 or 14.

16. The highlighting is to display the sparse region in a color different from a color of a region outside the sparse region.

15. The image display method according to claim 13 or 14.

17. 1. An image display system including a processor, The processor, Obtaining a plurality of idea vectors by vectorizing a plurality of idea texts based on input from a user; Displaying the plurality of idea texts at positions according to the components of the corresponding plurality of idea vectors; and generating a display image in which a plurality of reference texts are displayed at positions corresponding to respective components of a plurality of reference text vectors obtained by vectorizing a plurality of reference texts stored in advance in a memory; Displaying the display image on a display; the display image highlights a sparse region where the plurality of idea texts are sparse; Image display system.

Images