SNS article search support device and SNS article search support method

Abstract

This invention provides an SNS article search support device and SNS article search support method that can help users view SNS search results from diverse perspectives. [Solution] The article search support server includes a search query input module, an SNS search module that performs an SNS search and retrieves multiple SNS articles as search results, a classification axis query module that generates multiple classification axes for analyzing multiple SNS articles using a generation AI and retrieves the generated multiple classification axes from the generation AI, an article score query module that assigns scores to SNS articles according to the classification axes using a generation AI and retrieves the SNS articles to which scores have been assigned from the generation AI, and an article layout module that uses a classification axis selected from the multiple classification axes to create graph information in which nodes representing multiple SNS articles are placed at positions corresponding to the scores.

Description

Technical Field

[0001] The present disclosure relates to an SNS article search support device and an SNS article search support method.

Background Art

[0002] Regarding technologies for improving search accuracy and displaying useful attached information to users in Web and SNS (social networking service) searches, business scales are extremely large, and continuous technological development and patent applications are being carried out. Patent Document 1 discloses an information processing device (hereinafter referred to as "conventional device 1") aimed at identifying a search query that has no relevance to a predetermined event.

[0003] Conventional device 1 includes an acquisition unit that acquires reference distribution information indicating the distribution for each attribute of users who have performed actions related to a predetermined search query, a generation unit that generates target distribution information indicating the distribution for each attribute of users who have input the search query, a comparison unit that compares the reference distribution information and the target distribution information, and a specification unit that specifies a search query that satisfies a predetermined condition based on the comparison result by the comparison unit. Thereby, when noise is included in the user group to be analyzed, it is possible to identify a search query that has low relevance to the actions related to the predetermined search query from the search queries input by the user group, and thus it is possible to identify a search query that has no relevance to a predetermined event.

[0004] Patent Document 2 discloses a determination device (hereinafter referred to as "conventional device 2") aimed at providing an appropriate search service to users. Conventional device 2 includes an extraction unit that extracts tag candidates from a search target, and a determination unit that determines whether to adopt a tag candidate as a tag to be used at the time of search based on the tag candidates extracted by the extraction unit and the behavior history of users corresponding to the tag candidates. Conventional device 2 can provide an appropriate search service to users by determining valid tag candidates as tags for the search target based on the behavior history of users.

Prior Art Documents

[0005] [Patent Document 1] Japanese Patent Publication No. 2024-60955 [Patent Document 2] Patent No. 7520097 [Overview of the project] [Problems that the invention aims to solve]

[0006] While web searches also include searches for SNS information, individual SNS posts (sometimes referred to as "SNS articles") are short and, unlike general text, are difficult to understand in isolation, making it difficult to grasp the context of each search result. Generally, SNS algorithms analyze and learn from the search and click histories of individual internet users, resulting in a filter bubble where information that users want to see is prioritized, regardless of their preferences, and information that does not align with the user's perspective is isolated, creating an information environment where users become isolated within a "bubble" of their own thoughts and values. Furthermore, in SNS, when users express opinions within communities where they connect (follow) others with similar interests, they tend to receive only similar opinions in return (an echo chamber phenomenon), raising concerns that this narrows an individual's perspective and leads to a division of the world.

[0007] Therefore, SNS article search support devices are required to help users view SNS search results from diverse perspectives.

[0008] Conventional device 1 is characterized by identifying search queries that do not have any relevance to a predetermined event from among the search queries entered by a group of users. Conventional device 2 is characterized by determining valid tag candidates as search targets based on the user's behavior history. Conventional devices 1 and 2 do not propose multiple separation axes for users to evaluate SNS articles in search results, unlike the SNS article search support device of this disclosure.

[0009] This disclosure was made to solve the above-mentioned problems. Specifically, one of the purposes of this disclosure is to provide an SNS article search support device and an SNS article search support method that can help users view SNS search results from diverse perspectives. [Means for solving the problem]

[0010] To solve the above problems, the SNS article search support method of this disclosure is an article search support method using an information processing device, wherein the information processing device performs: a search query input into which a search query is entered; an SNS search that searches for SNS articles using the search query and obtains multiple SNS articles as search results; a classification axis query that inputs multiple SNS articles and a classification axis query prompt to a generation AI, causes the generation AI to generate multiple classification axes for analyzing the multiple SNS articles, and obtains the generated multiple classification axes from the generation AI; an article score query that inputs the SNS articles, the classification axes, and a score query prompt to the generation AI, causes the generation AI to assign scores to the SNS articles according to the classification axes, and obtains the SNS articles to which the scores have been assigned from the generation AI; and an article layout that uses a classification axis selected from the multiple classification axes to create graph information in which nodes representing multiple SNS articles are placed at positions corresponding to the scores.

[0011] The SNS article search support device of this disclosure is an SNS article search support device having an information processing device, the information processing device comprising: a search query input module into which a search query is input; an SNS search module that searches for SNS articles using the search query and obtains a plurality of SNS articles as search results; a classification axis query module that inputs a plurality of SNS articles and a classification axis query prompt to a generation AI, causes the generation AI to generate a plurality of classification axes for analyzing the plurality of SNS articles, and obtains the generated plurality of classification axes from the generation AI; an article score query module that inputs the SNS articles, the classification axes and a score query prompt to the generation AI, causes the generation AI to assign scores to the SNS articles according to the classification axes, and obtains the SNS articles to which the scores have been assigned from the generation AI; and an article layout module that uses a classification axis selected from the plurality of classification axes to create graph information in which nodes representing a plurality of SNS articles are placed at positions corresponding to the scores. [Effects of the Invention]

[0012] This disclosure helps users view social media search results from diverse perspectives. The effects described herein are not necessarily limited to those described herein and may include any of the effects described in this disclosure. [Brief explanation of the drawing]

[0013] [Figure 1] Figure 1 shows an example of a system configuration for an SNS article search system. [Figure 2] Figure 2 is a schematic diagram showing an example of the configuration of an article search support server. [Figure 3] Figure 3 shows an example of a conventional GUI screen. [Figure 4] Figure 4 is a sequence diagram illustrating an example of how the SNS article search system works. [Figure 5A] Figure 5A shows an example of the format of a classification axis query prompt. [Figure 5B] FIG. 5B is a diagram showing an example of the format of a scoring prompt. [Figure 6] FIG. 6 is a diagram showing an example of a GUI screen. [Figure 7] FIG. 7 is a diagram showing an example of a GUI screen. [Figure 8] FIG. 8 is a diagram for explaining an operation example of the SNS article search system. [Figure 9] FIG. 9 is a diagram showing an example of a GUI screen. [Figure 10] FIG. 10 is a diagram showing an example of a GUI screen. [Figure 11] FIG. 11 is a diagram showing an example of a GUI screen.

Mode for Carrying Out the Invention

[0014] Hereinafter, each embodiment of the present disclosure will be described with reference to the drawings. In all the drawings of the embodiments, the same or corresponding parts may be denoted by the same reference numerals.

[0015] > In the following description, various information may be represented by a data structure other than the described data structure. Further, in the following description, when the processing is described with a module as the subject, the subject of the processing may be a CPU or a device (server) instead of the module. The subject of the processing performed by executing the program may be an arithmetic unit, and may include a dedicated circuit for performing a specific processing. Here, the dedicated circuit is, for example, an FPGA (Field Programmable Gate Array), an ASIC (Application Specific Integrated Circuit), a CPLD (Complex Programmable Logic Device), or the like.

[0016] In the following description, the program may be installed from a program source into an information processing apparatus. The program source may be, for example, a program distribution server or a storage medium readable by the information processing apparatus. When the program source is a program distribution server, the program distribution server includes a processor and a storage resource for storing the program to be distributed, and the processor of the program distribution server may distribute the program to be distributed to other information processing apparatuses. Also, in each embodiment, two or more programs may be realized as one program, or one program may be realized as two or more programs.

[0017] <<First Embodiment>> An SNS article search system including an article search support server 100 according to the first embodiment will be described. FIG. 1 is a diagram showing a system configuration example of the SNS article search system. As shown in FIG. 1, the SNS article search system includes an article search support server 100, a search service providing server 200, a generation AI server 300, and a user terminal 400. The article search support server 100, the search service providing server 200, the generation AI server 300, and the user terminal 400 are communicably connected to each other via a network NW1. Note that a system including the article search support server 100, the search service providing server 200, and the generation AI server 300 may be referred to as an SNS article search system.

[0018] Figure 2 is a schematic diagram showing an example configuration of the article search support server 100. The article search support server 100 may be referred to as an "SNS article search support device". As shown in Figure 2, the article search support server 100 includes a non-volatile storage device 110 that can read and write data, memory 120 (e.g., RAM), a CPU 130, an input / output interface 140, a network interface 150, and a bus 160. For convenience, the device including the CPU 130, storage device 110, memory 120, input / output interface 140, network interface 150, and bus 160 is also referred to as an "information processing device". The information processing device may consist of multiple information processing devices, or it may be a virtual information processing device built on the cloud.

[0019] The storage device 110 includes (stores) a search query input module 111, an SNS search module 112, a classification axis query module 113, an article score query module 114, and an article layout module 115. Each of these modules is a set of programs executed by the CPU 130 to implement the functions of each module.

[0020] The search query input module 111 is a component into which a search query is entered. The SNS search module 112 is a component that retrieves multiple SNS articles as search results by performing a search for SNS articles using the search query. The classification axis query module 113 is a component that queries the generation AI server 300 for the classification axis when scoring the search results of SNS articles. The article score query module 114 is a component that queries the generation AI server 300 for a score for SNS articles according to the classification axis. The article layout module 115 is a component that creates a GUI (Graphical User Interface) screen for viewing the search results.

[0021] The CPU 130 loads the program stored in the storage device 110 into the memory 120. By executing the program loaded into the memory 120, the CPU 130 implements the various functions of the search query input module 111, the SNS search module 112, the classification axis query module 113, the article score query module 114, and the article layout module 115.

[0022] As mentioned above, the program to be executed by the CPU 130 is loaded into memory 120, and data used by the CPU 130 when executing the program is temporarily stored there.

[0023] The input / output interface 140 is an interface for connecting operating devices such as a keyboard and mouse, as well as a display. Search queries may also be entered via the operating devices connected to the input / output interface 140. The network interface 150 is an interface for connecting the article search support server 100 to the network NW1. Search queries may also be entered from the user terminal 400 to the article search support server 100 via the network NW1 connected to the network interface 150.

[0024] Furthermore, at least a portion of the processing performed by the CPU 130 when executing a program may be performed by other arithmetic units (e.g., hardware such as ASICs or FPGAs).

[0025] The search service provider server 200 has an SNS search API 210. The SNS search API 210 is an API (Application Programming Interface) that, in response to a request containing a search query, returns search results for SNS posts corresponding to the search query as a response.

[0026] The generation AI server 300 includes a large-scale language model API 310 and a language generation AI (not shown). The language generation AI includes large-scale language models (LLMs). The language generation AI is sometimes referred to as "generation AI". A large-scale language model (LLM) is a language model constructed using a large amount of data and deep learning technology. The large-scale language model API 310 is an API that, in response to a request, inputs information into the large-scale language model, generates an appropriate response, and returns the response as a response.

[0027] The search service server 200 and the generation AI server 300 are composed of the information processing equipment shown in Figure 2. The user terminal 400 is composed of the information processing equipment shown in Figure 2 and a display device (not shown) capable of displaying images connected to the input / output interface 140 of the information processing equipment. The user terminal 400 is, for example, a smartphone, tablet, or personal computer such as a notebook PC.

[0028] <Overview> Let me explain the overview. Generally, social media is considered problematic because it can lead to filter bubbles and echo chambers that narrow a person's perspective.

[0029] Therefore, as shown in Figure 3, the inventors of this application have realized a GUI that represents multiple SNS articles obtained as search results as nodes (points) and displays them on an interactive positioning map 3001 (see Non-Patent Literature "Journal Name: Interaction 2024 Proceedings, 2024 Information Processing Society of Japan, Paper ID 1B-29, P308-P313"). The interactive positioning map 3001 is an interactive positioning map (dialogue-type positioning map) that has the function of allowing the user to directly manipulate the graph. Note that a "positioning map" is a type of graph information and is a tool for visually showing how the evaluation target is positioned. A positioning map is typically represented as a two-dimensional graph, and different evaluation criteria are set for the X axis and Y axis.

[0030] This GUI linguistically analyzes SNS posts using two fixed axes: "Positive / Negative" and "Subjective / Objective." It assigns a score to each post, represents the SNS posts as nodes corresponding to their scores, and displays them on the interactive positioning map 3001.

[0031] This GUI displays an interactive positioning map 3001 with dynamic and interactive information next to the SNS post list (not shown in Figure 3), allowing users to refer to the interactive positioning map 3001 and prompt them to think, "Let's take a look at posts with a slightly different trend."

[0032] This GUI uses a fixed classification axis in the interactive positioning map 3001 for any search keyword, allowing users to view SNS posts from a fixed perspective. However, this GUI did not help users view SNS posts from multiple new perspectives.

[0033] In contrast, the inventors' SNS article search support device and method proposes multiple classification axes in real time, allowing users to view search results from multiple perspectives, regardless of the search keyword used. These proposed classification axes are not fixed in advance, but rather, they are suggested based on the multiple SNS articles in the search results, representing the content of the articles and promoting discussion. It is difficult and time-consuming for users to conceive and create multiple such classification axes based on the multiple SNS articles in the search results. In contrast, this SNS article search support device and method can automatically and quickly suggest the classification axis that best represents the content of the SNS articles and promotes discussion based on the multiple SNS articles in the search results.

[0034] This SNS article search support device / method displays an interactive positioning map, which is graph information with a classification axis selected by the user from among multiple classification axes, making it possible to visually display SNS articles scored according to that classification axis.

[0035] The classification axes of the interactive positioning map can be switched by the user from among the proposed classification axes, allowing users to view SNS posts from numerous new perspectives by switching classification axes.

[0036] Figure 4 is a sequence diagram illustrating an example of how the SNS article search system works.

[0037] Step S401: The user inputs a search query into the search query input module 111 via the user terminal 400.

[0038] S402: The SNS search module 112 obtains the search query from the search query input module 111 and sends the search query to the SNS search API 210 of the search service provider server 200.

[0039] S403: The search service provider server 200 sends the SNS search results (multiple SNS posts) to the SNS search module 112 as a response to the search query received by the SNS search API 210.

[0040] S404: The classification axis query module 113 retrieves SNS search results from the SNS search module 112 and sends the SNS search results list and the classification axis query prompt to the large-scale language model API 310.

[0041] Figure 5A shows the format of the classification axis query prompt. The classification axis query module 113 creates a classification axis query prompt by inputting the number of axes to be suggested in {axis_count}, the number of SNS posts included in the SNS search results list in {tweet_count}, and the SNS posts included in the SNS search results list to be analyzed in {tweets}. By using the classification axis query prompt format, the classification axis query module 113 can easily allow the generating AI to generate classification axes for SNS search posts. Note that the prompt created by the classification axis query prompt format in Figure 5A is merely an example. The prompt only needs to be an appropriate instruction to get the Large-Scale Language Model API to return appropriate classification axes for classifying the SNS posts to be specified, and is not limited to this example of description method.

[0042] S405: When the large-scale language model API 310 receives the SNS search result list and the classification axis query prompt, the language generation AI (large-scale language model) creates the number of classification axes (axis name and both poles of the axis (name of the first pole in the negative direction, name of the second pole in the positive direction)) requested by the SNS search result list and the classification axis query prompt, and sends the created classification axes to the classification axis query module 113.

[0043] S406: The article score query module 114 retrieves SNS articles and classification axes from the classification axis query module 113, and sends the SNS articles, classification axes, and score determination prompt to the large-scale language model API 310. Figure 5B shows the format of the score determination prompt. The article score query module 114 creates the score determination prompt by entering the names of both poles of each axis in {pole1} and {pole2}.

[0044] S407: When the large-scale language model API 310 obtains SNS posts, classification axes, and score determination prompts, the language generation AI (large-scale language model) assigns a score to the SNS posts according to the classification axes and sends it to the article score query module 114.

[0045] S408: The article layout module 115 obtains SNS articles and the scores assigned to them from the article score inquiry module 114, creates a GUI screen (image) using the SNS articles, the scores assigned to them, and other necessary data, and provides it to the user terminal 400.

[0046] Figure 6 shows an example of the GUI screen GM1 displayed on the display device of user terminal 400. Note that GUI screen GM1 is an example of the screen displayed when "presidential election" is entered as the search query (search keyword).

[0047] GUI screen GM1 includes a search display area R1, a search result evaluation area R2, and an article display area R3. Note that GUI screen GM1 in Figure 6 is displayed by the article layout module 115 (the same applies to GUI screens GM2 in Figure 7, GM3 in Figure 9, GM4 in Figure 10, and GM5 in Figure 11).

[0048] The search display area R1 includes a search box 601, a search button 602, first classification axis information 603a, second classification axis information 603b, and third classification axis information 603c. Hereafter, the first classification axis information 603a, second classification axis information 603b, and third classification axis information 603c will be referred to as "classification axis information 603" unless there is a need to distinguish between them.

[0049] The search box 601 is an input box where the search query (search keyword) is entered. In this example, "presidential election" is entered as the search query (search keyword) in the search box 601. The search button 602 is an operation button to start the search. The classification axis information 603 includes the name of the classification axis proposed for the SNS search results, the name of the first pole, and the name of the second pole.

[0050] The first classification axis information 603a includes "Media reporting stance" as the name of the classification axis, "Biased reporting" as the name of the first pole, and "Fair reporting" as the name of the second pole.

[0051] The second classification axis information 603b includes "Impact of the Presidential Election" as the name of the classification axis, "Negative Impact on the Economy" as the name of the first pole, and "Positive Impact on the Economy" as the name of the second pole.

[0052] The third classification axis information 603c includes "Political Leadership" as the name of the classification axis, "Irresponsible Statements" as the name of the first pole, and "Responsible Leadership" as the name of the second pole.

[0053] The search result evaluation area R2 includes an X-axis input box 611, a Y-axis input box 612, and an interactive positioning map 613.

[0054] The X-axis input box 611 contains a pull-down list showing the X-axis options for the interactive positioning map 613. The classification axes (axis names) included in this pull-down list correspond to the classification axes suggested for SNS search results. The user selects a classification axis to set as the X-axis from the options in the pull-down list by making a selection operation and inputs it into the X-axis input box 611.

[0055] The Y-axis input box 612 contains a pull-down list showing the Y-axis options for the interactive positioning map 613. The classification axes (axis names) included in this pull-down list correspond to the classification axes suggested for SNS search results. The user selects a classification axis to set as the Y-axis from the options in the pull-down list by making a selection operation and inputs it into the Y-axis input box 612.

[0056] In the interactive positioning map 613, nodes representing SNS articles are positioned on the interactive positioning map 613 according to the score assigned to the SNS article.

[0057] By hovering the cursor (not shown) over a specific node in the interactive positioning map 613 and clicking, users can scroll and display detailed information about SNS articles in the article display area R3. The article search support server 100 maintains a management table that manages the correspondence between nodes on the interactive positioning map 613 and SNS articles in the search results. The article search support server 100 refers to the management table and scrolls and displays detailed information about the SNS article corresponding to the manipulated node in the article display area R3. The interactive positioning map 613 provides more information than a static graph and is a powerful tool to help users understand the evaluation of each SNS article. The article search support server 100 helps users understand the evaluation of each SNS article from multiple classification axes (perspectives) proposed.

[0058] The article display area R3 contains multiple articles 614 from the search results. In the article display area R3, among the multiple articles 614, the SNS article corresponding to a specific node selected by the user is displayed in a different format than the other SNS articles. In the example in Figure 6, the node filled in black represents the SNS article selected by the user. In the article display area R3, detailed information about the SNS article corresponding to the selected node is displayed in a different format than the other SNS articles (in this example, the area around the SNS article's display area is filled in black).

[0059] When a user enters a search keyword (search query) for SNS search into the search box 601 on the GUI screen GM1 of the user terminal 400, the search results matching the search query are collected via the article search support server 100 and fed into a large-scale language model to propose multiple classification axes. For example, for a search query such as "President search," the system proposes three classification axes (first pole / second pole): media reporting stance (biased reporting / fair reporting), impact of presidential elections (negative impact on the economy / positive impact on the economy), and political leadership (irresponsible statements / responsible leadership).

[0060] The article search support server 100 inputs each of the proposed classification axes (first pole / second pole) and the search results into a large-scale language model, and assigns scores to each of the multiple classification axes (first pole / second pole) for the numerous SNS articles in the search results.

[0061] On the GUI screen GM1, the user selects an axis for the interactive positioning map 613 from the three proposed classification axes, and displays the interactive positioning map 613 with the selected axis.

[0062] The interactive positioning map 613 displays nodes representing scored articles, with each node positioned according to its score. The user interacts with the nodes on the interactive positioning map 613 via the user terminal 400 (for example, by hovering the cursor over a node and clicking), causing the details of the SNS article corresponding to the manipulated node to scroll and be displayed in the article display area R3.

[0063] The user can also display the interactive positioning map 613 with a changed axis. Figure 7 shows an example of GUI screen GM2 after switching the X-axis of the interactive positioning map 613 in GUI screen GM1 of Figure 6.

[0064] On the GUI screen GM1, the user re-selects a classification axis for the interactive positioning map 613 from the three proposed classification axes, and displays the interactive positioning map 613 with the re-selected classification axis. For example, on the GUI screen GM1 in Figure 6, the user inputs "Media coverage stance" into the X-axis input box 611 by selecting "Media coverage stance" from the options in the pull-down list of the X-axis input box 611. As a result, the interactive positioning map 613 in Figure 6 switches to the interactive positioning map 613 in Figure 7.

[0065] In the interactive positioning map 613 shown in Figure 7, the X-axis represents "media reporting stance," with the first pole of the X-axis representing "biased reporting" and the second pole representing "fair reporting." Thus, the X-axis has been switched from the X-axis in Figure 6.

[0066] The Y-axis represents the "impact of the presidential election," with the first pole of the Y-axis representing "negative impacts on the economy" and the second pole representing "positive impacts on the economy." Note that the Y-axis is the same as the Y-axis in Figure 6.

[0067] In the interactive positioning map 613, nodes representing SNS articles scored according to the re-selected classification axis are rearranged to positions corresponding to their scores. The user can view articles from a new perspective on the search results using the interactive positioning map 613 after the classification axis has been switched. In this way, the user can view articles from various perspectives by switching between the multiple proposed classification axes.

[0068] <Effects> As described above, the SNS article search support device according to the first embodiment can help users view SNS search results from a variety of perspectives.

[0069] <<Second Embodiment>> A social networking service (SNS) article search system including an article search support server 100 (SNS article search support device) according to the second embodiment will be described. The article search support server 100 according to the second embodiment differs from the article search support server 100 according to the first embodiment only in the following respects.

[0070] The article search support server 100 according to the second embodiment allows users to add new classification axes and can display a positioning map with the added classification axes.

[0071] The following explanation will focus on these differences.

[0072] Figure 8 is a sequence diagram illustrating the operation of an SNS article search system including an article search support server 100 according to the second embodiment.

[0073] Sections S401 through S408 have already been described, so their explanation will be omitted.

[0074] S409: The user inputs a new classification axis devised by the user into the article score query module 114 via the user terminal 400.

[0075] S410: The article score query module 114 obtains the name of the classification axis devised by the user via the user terminal 400, and then sends the name of the classification axis, a prompt, and the search results articles to the large-scale language model API 310. The prompt includes a request to the large-scale language model to name the first and second poles of the classification axis, and a request to score the SNS articles in the search results according to the new classification axis.

[0076] S411: The large-scale language model API 310 uses a language generation AI (large-scale language model) to assign names to the first and second poles of a classification axis, assigns a score to SNS articles according to the classification axis, and sends it to the article score query module 114.

[0077] S412: The article layout module 115 obtains SNS articles and the scores assigned to them from the article score inquiry module 114, creates a GUI screen (image) using the SNS articles, the scores assigned to them, and other necessary data, and provides it to the user terminal 400.

[0078] Figure 9 is a diagram illustrating an example of the GUI screen GM3. The GUI screen GM3 includes an axis addition input box 911 and an add button 912 in the search result evaluation area R2. Except for this point, it is the same as the GUI screen GM2 in Figure 7 of the first embodiment.

[0079] The axis addition input box 911 is an input box where the name of the classification axis is entered. Names of classification axes other than the proposed classification axis are entered in this axis addition input box 911. The add button 912 is operated to add the classification axis entered in the axis addition input box 911 to the respective dropdown lists for the X-axis and Y-axis.

[0080] Users can also add their own classification axes and display an interactive positioning map 613 that incorporates those classification axes.

[0081] For example, as shown in GUI screen GM4 in Figure 10, the user enters the name of the classification axis they have chosen, "Entertainment Elements," into the axis addition input box 911 on GUI screen GM3 in Figure 9. The user then operates the add button 912. As a result, the classification axis "Entertainment Elements" is added to the options in both the dropdown list showing the X-axis options and the dropdown list showing the Y-axis options.

[0082] Furthermore, as shown in the GUI screen GM5 in Figure 11, the user inputs "entertainment elements" into the Y-axis input box 612 by selecting "entertainment elements" from the options in the pull-down list of the Y-axis input box 612. As a result, the interactive positioning map 613 in Figure 10 switches to the interactive positioning map 613 in Figure 11.

[0083] The user can view articles based on newly conceived and entered classification axes for the search results using the switched-to interactive positioning map 613. In this way, the user can view articles from various perspectives by switching between the suggested classification axes and the classification axes newly conceived by the user.

[0084] <Effects> As described above, the SNS article search support device according to the second embodiment can help users view SNS search results from diverse perspectives. Furthermore, the SNS article search support device according to the second embodiment can help users view SNS search results from perspectives that they themselves have considered and added.

[0085] <<Variation>> This disclosure is not limited to the embodiments described above, and various modifications can be adopted within the scope of this disclosure. Furthermore, the embodiments described above can be combined with each other without departing from the scope of this disclosure. In each of the embodiments described above, the interactive positioning map 613 has two axes, but it may have three or more. In each of the embodiments described above, the article search support server 100 may be equipped with a language generation AI. In this case, the classification axis query module 113 may query the language generation AI of the article search support server 100 instead of the language generation AI of the generation AI server 300. The article score query module 114 may query the language generation AI of the article search support server 100 instead of the language generation AI of the generation AI server 300.

[0086] In each of the above embodiments, the screen layout, such as that shown in Figure 6, is not limited to that shown in each embodiment. The search method (search box 601 and search button 602), the article display area (R3), the display area of ​​the interactive positioning map 613 (R2), etc., may be arranged differently from those in this embodiment. Furthermore, in each of the above embodiments, the search method (search box 601 and search button 602), the article display area (R3), the display area of ​​the interactive positioning map 613 (R2), etc., do not necessarily have to be displayed on the same screen. They may be displayed on screens that the user can easily switch between. In each of the above embodiments, the nodes corresponding to each article in the interactive positioning map 613 may be circular or other shapes, and may use illustrations or animations.

[0087] In each of the above embodiments, the user terminal 400 (information processing device included in the user terminal 400) may include a search query input module 111, an SNS search module 112, a classification axis query module 113, an article score query module 114, and an article layout module 115. In this case, the SNS article search system in Figure 1 does not need to include an article search support server 100, and the operation of the article search support server 100 described above may be performed by the user terminal 400. [Explanation of Symbols]

[0088] 100... Article search support server, 111... Search query input module, 112... SNS search module, 113... Classification axis query module, 114... Article score query module, 115... Article layout module, 200... Search service provider server, 300... Generative AI server, 310... Large-scale language model API, 400... User terminal, 611... X-axis input box, 612... Y-axis input box, 613... Interactive positioning map, 911... Axis addition input box, 912... Add button, GM1~GM5... GUI screens

Claims

1. An SNS article search support device having an information processing device, The aforementioned information processing device is A search query input module where the search query is entered, An SNS search module that searches for SNS posts using the aforementioned search query and retrieves multiple SNS posts as search results, A classification axis query module that inputs multiple SNS articles and classification axis query prompts to a generating AI, causes the generating AI to generate multiple classification axes for analyzing the multiple SNS articles, and retrieves the generated multiple classification axes from the generating AI. An article score inquiry module that inputs the aforementioned SNS article, the classification axis, and the score inquiry prompt to the generating AI, causes the generating AI to assign a score to the SNS article according to the classification axis, and retrieves the SNS article to which the score has been assigned from the generating AI. An article layout module that creates graph information by using the classification axis selected from among the multiple classification axes and arranging nodes representing multiple SNS articles at positions corresponding to the scores, Having, SNS article search support device.

2. In the SNS article search support device according to claim 1, The aforementioned graph information is an interactive positioning map in which the nodes can be manipulated. The aforementioned article layout module is When the node on the interactive positioning map is manipulated, a GUI screen is created in which the SNS article corresponding to the manipulated node is displayed in a specific display area. SNS article search support device.

3. In the SNS article search support device according to claim 2, The aforementioned article layout module is An X-axis input box is operated to select and input the classification axis to be set as the X-axis of the interactive positioning map from among a plurality of classification axes obtained by the separation axis inquiry module, A Y-axis input box is operated to select and input the classification axis to be set as the Y-axis of the interactive positioning map from among a plurality of classification axes obtained by the separation axis query module, The interactive positioning map has the classification axis entered in the X-axis input box on the X-axis and the classification axis entered in the Y-axis input box on the Y-axis, To create the GUI screen that includes the above, SNS article search support device.

4. In the SNS article search support device according to claim 3, The aforementioned article layout module is The GUI screen is created which further includes a search article display area that includes multiple SNS articles from the search results. SNS article search support device.

5. In the SNS article search support device according to claim 4, The aforementioned article layout module is When the node on the interactive positioning map is operated, the GUI screen is created that displays the SNS article corresponding to the operated node in a different display format than other SNS articles. SNS article search support device.

6. In the SNS article search support device according to claim 4, The aforementioned article layout module is A search input box that is operated on to enter the aforementioned search query, The search button, which is used to start a search, To create the GUI screen that includes the above, SNS article search support device.

7. In the SNS article search support device according to claim 2, The aforementioned article layout module is An X-axis input box is operated to select and input the classification axis to be set as the X-axis of the interactive positioning map from among a plurality of classification axes obtained by the separation axis inquiry module, A Y-axis input box is operated to select and input the classification axis to be set as the Y-axis of the interactive positioning map from among a plurality of classification axes obtained by the separation axis query module, The interactive positioning map has the classification axis entered in the X-axis input box on the X-axis and the classification axis entered in the Y-axis input box on the Y-axis, A new classification axis is entered in the axis addition input box, An additional button to enable input of the new classification axis entered into the axis addition input box into the X-axis input box and the Y-axis input box, Create the GUI screen that includes the above, The aforementioned article's score query module is, By inputting the aforementioned SNS article, the name of the classification axis entered in the axis addition input box, and the prompt into the generating AI, the generating AI is made to create both poles of the classification axis, assign the score to the SNS article according to the classification axis, and retrieve the SNS article to which the score has been assigned from the generating AI. SNS article search support device.

8. A method for supporting SNS article searches using an information processing device, The aforementioned information processing device The search query input field is where the search query is entered, An SNS search that performs a search for SNS posts using the aforementioned search query and retrieves multiple such SNS posts as search results, A classification axis query involves inputting multiple SNS articles and classification axis query prompts into a generating AI, causing the generating AI to generate multiple classification axes for analyzing the multiple SNS articles, and obtaining the generated multiple classification axes from the generating AI. By inputting the aforementioned SNS article, the classification axis, and the score inquiry prompt into the generating AI, the generating AI assigns a score to the SNS article according to the classification axis, and the article score inquiry retrieves the SNS article to which the score has been assigned from the generating AI. An article layout that creates graph information by using the classification axis selected from among the multiple classification axes and arranging nodes representing multiple SNS articles at positions corresponding to the scores, To do Methods for assisting with SNS article searches.

Images