Information processing method, information processing program, and information processing apparatus

Abstract

To provide an information processing method, an information processing program, and an information processing apparatus configured to create and output comments that specialists would likely make for an unknown or new case.SOLUTION: An information processing method causes a computer to execute the processes of: acquiring an inquiry including the content of activity; acquiring, using a learning model, a comment on biodiversity for the acquired inquiry; and outputting the acquired comment.SELECTED DRAWING: Figure 13

Description

Technical Field

[0001] The present invention relates to an information processing method, an information processing program, and an information processing apparatus that output comments.

Background Art

[0002] In recent years, in order to conserve biodiversity, the idea that certain restrictions should be placed on human activities has been spreading. In particular, in development projects carried out by companies, depending on the content of the project and the location where the project is implemented, the impact on the environment can be significant. In this case, from the perspective of biodiversity conservation, it may be necessary to abandon the implementation of the project.

[0003] Conventionally, in order to determine the impact of project implementation, after formulating a project plan, methods such as environmental assessment have been used to predict the impact on biodiversity. However, such work requires costs and time, and if the project is abandoned, it will result in losses for the company. Therefore, there is a desire to collect opinions from academia and determine the pros and cons of the plan at an early stage of project plan formulation. In relation to such a situation, Patent Document 1 proposes a system for collecting expert opinions on articles.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, in the conventional system, for unknown or new cases without accumulation, time is required to collect opinions. The present invention has been made in view of such a situation. Its purpose is to provide an information processing method, an information processing program, and an information processing apparatus that create and output comments that an expert would make for unknown or new cases.

Means for Solving the Problem

[0006] An information processing method according to an aspect of the present application acquires an inquiry including the content of an activity, acquires a comment on biodiversity for the acquired inquiry using a learning model, and a computer executes a process of outputting the acquired comment.

Effect of the Invention

[0007] In one aspect of the present application, it becomes possible to output a comment that an expert would make.

Brief Description of the Drawings

[0008]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Figure 19

Figure 20

Figure 21

Figure 22

Figure 23

Mode for Carrying Out the Invention

[0009] The following embodiments will be described with reference to the drawings. In the following description, "academia" mainly refers to researchers in academic research institutions and the like. It should be noted that although it is interpreted that academia does not include researchers belonging to a company that has a stake in a specific company, here, researchers in private research institutions are also included in academia. "Activity" mainly has in mind the business activities carried out by for-profit enterprises, but also includes social activities carried out by non-profit organizations, groups without legal personality, and individuals. In the following description, an activity is an activity that affects or may affect biodiversity. A business plan is a plan for business activities and includes the content of the business. It is desirable that the business plan includes the location of the business.

[0010] FIG. 1 is an explanatory diagram showing a configuration example of a comment system. The comment system 100 includes a comment server 1, a corporate terminal 2, and an academic terminal 3. The comment server 1, the corporate terminal 2, and the academic terminal 3 are communicably connected to each other via a network N. Further, an interaction system 4 is available via the network N for the comment system 100. In FIG. 1, one corporate terminal 2 and one academic terminal 3 are shown, but each may be two or more.

[0011] The comment server 1 outputs comments on the input business plan. The comment server 1 is composed of a server computer, a workstation, a PC (Personal Computer), etc. Further, the comment server 1 may be composed of a multi-computer consisting of a plurality of computers, a virtual machine virtually constructed by software, or a quantum computer. Furthermore, the function of the comment server 1 may be realized by a cloud service.

[0012] FIG. 2 is a block diagram showing a hardware configuration example of the comment server. The comment server 1 includes a control unit 11, a main storage unit 12, an auxiliary storage unit 13, a communication unit 14, and a reading unit 15. The control unit 11, the main storage unit 12, the auxiliary storage unit 13, the communication unit 14, and the reading unit 15 are connected by a bus B.

[0013] The control unit 11 has an arithmetic processing device such as one or more CPUs (Central Processing Unit), MPUs (Micro-Processing Unit), GPUs (Graphics Processing Unit). The control unit 11 reads and executes a control program 1P (program, program product) stored in the auxiliary storage unit 13 to perform various information processes, control processes, etc., and realizes various functional units.

[0014] The main memory unit 12 is SRAM (Static Random Access Memory), DRAM (Dynamic Random Access Memory), flash memory, etc. The main memory unit 12 mainly temporarily stores data necessary for the control unit 11 to execute arithmetic processing.

[0015] The auxiliary storage unit 13 is a hard disk or an SSD (Solid State Drive), etc., and stores the control program 1P and various DBs (Databases) necessary for the control unit 11 to execute processing. The auxiliary storage unit 13 stores the enterprise DB 131, the academy DB 132, the thesaurus DB 133, the article DB 134, the comment DB 135, and the knowledge DB 136. The auxiliary storage unit 13 may be a separate body from the comment server 1 and may be an externally connected external storage device. Various DBs, etc. stored in the auxiliary storage unit 13 may be stored in a database server or cloud storage different from the comment server 1.

[0016] The communication unit 14 communicates with the enterprise terminal 2, the academy terminal 3, and the dialogue system 4 via the network N. Also, the control unit 11 may use the communication unit 14 to download the control program 1P from another computer via the network N or the like and store it in the auxiliary storage unit 13.

[0017] The reading unit 15 reads a portable storage medium 1a including a CD (Compact Disc)-ROM and a DVD (Digital Versatile Disc)-ROM. The control unit 11 may read the control program 1P from the portable storage medium 1a via the reading unit 15 and store it in the auxiliary storage unit 13. Also, the control unit 11 may read the control program 1P from the semiconductor memory 1b.

[0018] The enterprise terminal 2 is a terminal used by the person in charge of the enterprise who requests comments on the business plan. The enterprise terminal 2 is composed of a notebook computer, a panel computer, a tablet computer, a smartphone, etc.

[0019] FIG. 3 is a block diagram showing an example of the hardware configuration of the enterprise terminal. The enterprise terminal 2 includes a control unit 21, a main memory unit 22, an auxiliary storage unit 23, a communication unit 24, an input unit 25, and a display unit 26. Each component is connected by a bus B.

[0020] The control unit 21 has one or more arithmetic processing devices such as a CPU, MPU, and GPU. The control unit 21 provides various functions by reading and executing the control program 2P (program, program product) stored in the auxiliary storage unit 23.

[0021] The main memory unit 22 is an SRAM, DRAM, flash memory, etc. The main memory unit 22 mainly temporarily stores data necessary for the control unit 21 to execute arithmetic processing.

[0022] The auxiliary storage unit 23 is a hard disk or SSD, etc., and stores various data necessary for the control unit 21 to execute processing. The auxiliary storage unit 23 may be an external storage device connected to the enterprise terminal 2. Various DBs, etc. stored in the auxiliary storage unit 23 may be stored in a database server or cloud storage.

[0023] The communication unit 24 communicates with the comment server 1 via the network N. Also, the control unit 21 may use the communication unit 24 to download the control program 2P from another computer via the network N or the like and store it in the auxiliary storage unit 23.

[0024] The input unit 25 is a keyboard or a mouse. The display unit 26 includes a liquid crystal display panel or an organic EL (electro Luminescence) display panel, etc. The display unit 26 displays comments output by the comment server 1. Also, the input unit 25 and the display unit 26 may be integrated to form a touch panel display. Note that the enterprise terminal 2 may display on an external display device.

[0025] The academic terminal 3 is a terminal used by academia. The academic terminal 3 is composed of a notebook computer, a panel computer, a tablet computer, a smartphone, etc.

[0026] Figure 4 is a block diagram showing an example of the hardware configuration of the academic terminal. The academic terminal 3 includes a control unit 31, a main memory unit 32, an auxiliary storage unit 33, a communication unit 34, an input unit 35, and a display unit 36. Each component is connected by a bus B. Since the control unit 31, the main memory unit 32, the auxiliary storage unit 33, the communication unit 34, the input unit 35, and the display unit 36 have the same configurations as the control unit 21, the main memory unit 22, the auxiliary storage unit 23, the communication unit 24, the input unit 25, and the display unit 26 of the enterprise terminal 2 respectively, the description is omitted.

[0027] The dialogue system 4 includes a dialogue server 41 and an LLM 42 (learning model). The dialogue system 4 provides an interactive AI (Artificial Intelligence) service using the LLM 42 (Large Language Models) under the control of the dialogue server 41. The LLM 42 is a natural language processing model trained using a large amount of text data. The dialogue system 4 is configured using Transformer, BERT, GPT, ChatGPT, BARD, etc. Note that the comment server 1 may be equipped with a large language model and the comment server 1 may provide an interactive AI service.

[0028] Next, the database will be described. Figure 5 is an explanatory diagram showing an example of the enterprise DB. The enterprise DB 131 stores information about enterprises. The enterprise DB 131 includes an enterprise ID column, a name column, and an industry column. The enterprise ID column stores the enterprise ID that identifies the enterprise. The name column stores the name of the enterprise. The industry column stores the industry of the business that the enterprise is engaged in.

[0029] FIG. 6 is an explanatory diagram showing an example of an Academia DB. The Academia DB 132 stores information about academia. The Academia DB 132 includes an AID column, a name column, an affiliation category column, an affiliation column, a title column, and a field of expertise column. The AID column stores an AID that can uniquely identify an academia. The name column stores the name of the academia. The affiliation category column stores the category of the research institution to which the academia belongs. The category is a university or a national public research institution, etc. The affiliation column stores the name of the research institution to which the academia belongs. The title column stores the title of the academia. The field of expertise column stores the field of expertise of the academia. The academia registered in the Academia DB 132 is a pre-specified expert who can comment on biodiversity.

[0030] FIG. 7 is an explanatory diagram showing an example of a Thesaurus DB. The Thesaurus DB 133 stores thesaurus information. The Thesaurus DB 133 includes a thesaurus ID column, a representative name column, a thesaurus column, an associative word column, and a supplementary note column. The thesaurus ID column stores a thesaurus ID that identifies a thesaurus. The representative name column stores a word used when unifying notations. The thesaurus column stores a plurality of words in a similar relationship. The associative word column stores words associated from the representative name and the thesaurus. The supplementary note column stores supplementary matters and memos.

[0031] FIG. 8 is an explanatory diagram showing an example of an article DB. The article DB 134 stores news articles such as newspapers and magazines. The article DB 134 includes an article ID column, a title column, a body column, a location column, an author column, a reference column, a distribution date / time column, a deletion date / time column, a correction date / time column, a search index column, a main phrase column, a category column, a sentence vector column, and a summary column. The article ID column stores an article ID that identifies an article. The title column stores the title of the article. The body column stores the body of the article. The location column stores the location where the activity mentioned in the article took place. Here, the location is not limited to information indicating a geographical position, and information on locations with different viewpoints when considering the influence on biodiversity due to different natural environments such as a sandy beach, a riverbank, a wetland, etc. may be stored. Also, as locations, proper names of locations where the natural environment can be grasped such as Tokyo Bay, Kasumigaura, Kamikochi, etc. may be stored. The determination of the location is performed, for example, by performing syntax analysis of the article body, extracting candidate words, and comparing whether the extracted words are words indicating a location with a pre-prepared dictionary. The author column stores the author's name when the article is a signed article. The reference column stores the material information if there is a reference for the article. The distribution date / time column stores the date and time when the article was distributed. The deletion date / time column stores the date and time when the article has already been deleted. The correction date / time column stores the date and time when the content of the article was corrected after distribution. The search index column stores words that serve as indexes when searching for articles. The main phrase column stores the main phrases that make up the article content. The category column stores the category of the article. The sentence vector column stores a feature vector representing the content of the article. The summary column stores a summary of the article represented by three sentences.

[0032] FIG. 9 is an explanatory diagram showing an example of a comment DB. The comment DB 135 stores the comments of the academy on the article. The comment DB 135 includes a comment ID column, an article ID column, a text column, an AID column, a reference column, a creation date / time column, a deletion date / time column, a correction date / time column, a correction destination ID column, an evaluation column, and a parent ID column. The comment ID column stores the comment ID that identifies the comment. The article ID column stores the article ID of the article that is the target of the comment. The text column stores the text of the comment. The AID column stores the AID of the academy that made the comment. The reference column stores information about the material if there is a material related to the comment. The creation date / time column stores the date and time when the comment was created. The deletion date / time column stores the date and time when the comment was deleted if the comment has been deleted. The correction date / time column stores the correction date and time if the comment has been corrected. The correction destination ID column stores the comment ID of the comment after correction. The evaluation column stores the evaluation of the comment. When the evaluation is stored as a numerical value, the evaluation value takes a five-level value from 1 to 5, for example, with 5 being the highest evaluation and 1 being the lowest evaluation. The evaluation may be represented by symbols such as a circle, a triangle, a cross, or an image indicating "like". The parent ID column stores the comment ID of the comment to which the comment is attached if the comment is a comment on another comment.

[0033] FIG. 10 is an explanatory diagram showing an example of a knowledge DB. The knowledge DB 136 stores by associating article content with academic comments thereon. The knowledge DB 136 includes a data ID column, an article ID column, an activity content column, an activity location column, an article vector column, a comment column, a comment vector column, an AID column, a distribution date and time column, a search index column, a main phrase column, and a category column. The data ID column stores a data ID for specifying training data. The article ID column stores the article ID of a related article. The activity content column stores the activity content included in the article. The activity content is, for example, the business content of a company. The activity location column stores information on the location where the activity content included in the article is performed. The information on the location is the value of the location column of the article DB 134. The article vector column stores a feature vector generated from the article. The comment column stores a comment on the article, that is, the content of the text column of the comment DB 135. The comment vector column stores a feature vector generated from the comment. The AID column stores the AID of the academy that made the comment. The distribution date and time column stores the date and time when the article was distributed. The search index column stores the search index assigned to the article. The main phrase column stores the main phrases constituting the article. The category column stores the category of the activity content.

[0034] The above-described feature vectors are vectors in natural language processing. Since the process of vectorizing a text is a known technique, the description thereof is omitted.

[0035] Next, the information processing performed by the comment system 100 will be described. FIG. 11 is a flowchart showing an example of the procedure for request processing. The request processing is a process of requesting comments on the article information from academia. The request processing is periodically activated as a batch process, or is activated when a robot program such as a crawler detects that article information that can be a collection target has been distributed. The control unit 11 of the comment server 1 collects articles (step S1). The control unit 11 collects articles from a preset news site or the like. When articles have been collected by a crawler or the like, the control unit 11 collects the articles from the crawler or the like. The control unit 11 selects one article to be processed (step S2). The control unit 11 creates article data (step S3). The control unit 11 creates an extraction of the title, a selection of main phrases, an assignment of a category, a creation of a feature vector, a summary, etc. Since these processes can be realized by known techniques, the description is omitted. Note that a part of the creation work may be performed manually. The control unit 11 stores the article data in the article DB 134 (step S4). The control unit 11 creates and edits a request for comments on the article (step S5). The control unit 11 selects the academia to be requested as necessary. Although it is assumed that requests are made to all academias, depending on the category of the article, requests are made only to academias specialized in the preset academic fields. When a request to the academia to be requested has already been created, the control unit 11 edits the request and adds the request content. When the request has not yet been created, the control unit 11 newly creates a request. The control unit 11 determines whether there is unprocessed article data (step S6). When the control unit 11 determines that there is unprocessed article data (YES in step S6), the process returns to step S2, and the unprocessed article data is processed. When the control unit 11 determines that there is no unprocessed article data (NO in step S6), the control unit 11 sends a request for creating comments to the academia (step S7) and ends the process. The transmission means is an e-mail or a push notification or the like.

[0036] FIG. 12 is a flowchart showing an example of a generation process. The generation process is a process of generating knowledge data in which article information and comments from academia for the article are associated with each other. The generation process is activated, for example, when an academy selects a hyperlink included in a comment request. The control unit 31 of the academy terminal 3 transmits a request for a comment input form to the comment server 1 (step S11). The control unit 11 of the comment server 1 receives the request (step S12). The control unit 11 determines whether comments for the target article are being accepted (step S13). When the control unit 11 determines that comments are being accepted (YES in step S13), it creates an input screen for comments (step S14). The control unit 11 transmits the created input screen to the academy terminal 3 (step S15). The control unit 31 of the academy terminal 3 receives and displays the input screen (step S16). The academy inputs comments into the input screen and gives a transmission instruction. The control unit 31 accepts the comments and transmits them to the comment server 1 (step S17). The control unit 11 of the comment server 1 receives the comments (step S18). The control unit 11 generates knowledge data in which the article information and the comments are associated with each other (step S19). The control unit 11 stores the knowledge data in the knowledge DB 136 (step S20). The control unit 11 transmits the completion of acceptance to the academy terminal 3 (step S21). The control unit 31 of the academy terminal 3 receives and displays the completion (step S22). The control unit 31 ends the process. When the control unit 11 determines that comments are not being accepted (NO in step 13), it transmits a message indicating that comments cannot be accepted to the academy terminal 3 (step S23). The control unit 31 of the academy terminal 3 receives the message (step S24). The control unit 31 displays the message (step S25) and ends the process.

[0037] Note that whether comments are being accepted is stored in the article DB 134 or the like. Comments are accepted for a predetermined period after the article is distributed. Or, comments are accepted until the number of comments reaches a predetermined number. Regardless of the period or the number of comments, comments may be accepted until the target article is deleted.

[0038] FIG. 13 is a flowchart showing an example of the procedure of comment generation processing. The comment generation processing is a process of generating comments that would be sent from academia for a company's new business or the like using knowledge data. A person in charge at the company inputs a business plan including the content of the business and the location where the business is to be carried out into the company terminal 2. The control unit 21 of the company terminal 2 receives the business plan (step S41). When the person in charge operates a button to request comments, the control unit 21 transmits the received business plan to the comment server 1 (step S42). The control unit 11 of the comment server 1 receives the business plan (step S43). The control unit 11 vectorizes the text indicating the content of the business included in the business plan (step S44). At this time, the control unit 11 refers to the thesaurus DB 133 and replaces the terms used with representative names, etc. Then, the control unit 11 creates a feature vector from the text. The control unit 11 uses the created feature vector to refer to the knowledge DB 136 and extracts articles with similar content and comments on them (step S45). For the similarity determination, Euclidean distance, cosine similarity, etc. are used. The control unit 11 sets a role (step S46). The control unit 11 transmits a prompt including the role, the inquiry about the business plan, and the extracted comments to the dialogue system 4 (step S47). The role is a text that instructs the behavior to the dialogue system 4. For example, "You are an editor who collects comments on the impact of corporate activities on biodiversity from multiple academic institutions and creates representative comments on them." is set as the role. The dialogue system 4 transmits the role, the inquiry about the business plan, and the answer to the comment to the comment server 1. The number of comments to be created by the dialogue system 4 may be specified. The control unit 11 of the comment server 1 receives the answer (step S48). The control unit 11 creates a screen including the received answer (step S49). The control unit 11 transmits the created screen to the company terminal 2 (step S50). The control unit 21 of the company terminal 2 receives the screen (step S51). The control unit 21 displays the screen (step S52) and ends the process.

[0039] (Example of prompt) Examples of prompts sent to the dialogue system 4 in the comment generation process are shown below.

[0040] You are an editor who collects comments on the impact of corporate activities on biodiversity from multiple academic institutions and creates representative comments from them. {Content of the user's inquiry} Please refer to the following article information when creating. {Article ID ···} {Article ID ···} {Article ID ···} Please list the researcher's name, comment, and evaluation value separately for each researcher in the output format.

[0041] When sending the above prompt, the control unit 11 creates a virtual table (View) including the article ID, the text of the article, the comment ID, the name of the researcher who made the comment, the text of the comment, and the evaluation of the comment based on the academic institution DB 132, the article DB 134, and the comment DB 135, and sets it to be referable by the dialogue system 4.

[0042] (Narrowing down the target article) In the comment generation process, in the extraction of articles containing content similar to the content of the business, it is assumed that the similarity is calculated by brute force. However, since the computational complexity may become extremely large if it is brute force, the target articles may be narrowed down. For example, if the content of the business input by the person in charge includes location information, only articles having location information of the same or similar locations are targeted for extraction.

[0043] Also, the attributes of comments on the article may be used as the filtering conditions. Conditions may include the degrees obtained by the academia that made the comments (master's or below, doctorate), and the affiliation of the academia (university, national public research institution, private research institution). It is also possible to filter by the creation period of the comments. Select the comments that meet the conditions, and use the articles associated with the selected comments as the extraction targets. In addition, comments may be selected based on the evaluation values of the comments described later and the evaluation values of the academia that made the comments, and the articles associated with the selected comments may be used as the extraction targets. Note that the conditions for comments are also used when selecting comments associated with similar articles.

[0044] The process of filtering comments is performed after extracting the comments in step S45 of FIG. 13. In the case of conditions regarding academia, the control unit 11 of the comment server 1 acquires the AID of the academia that created the extracted comments. The control unit 11 searches the academia DB 132 using the acquired AID as the search key to obtain the information of the academia. The control unit 11 compares the information of the acquired academia with the filtering conditions. If the filtering conditions are met, the comment is left as a processing target; if the filtering conditions are not met, the comment is excluded from the processing target. In the case of conditions regarding comments, the control unit ******** compares the attributes (creation date, evaluation, etc.) of the extracted comments with the filtering conditions. If the filtering conditions are met, the comment is left as a processing target; if the filtering conditions are not met, the comment is excluded from the processing target.

[0045] (Modification example) In the comment creation process of FIG. 13, a plurality of past articles similar to the input business content and the comments thereon are sent to the dialogue system 4 to generate comments, but it is not limited to this. Fine-tuning may be performed on the LLM 42 of the dialogue system 4, and the input business content may be sent to the dialogue system 4 to receive comments thereon.

[0046] FIG. 14 is a flowchart showing an example of the procedure for executing fine-tuning. The control unit 11 of the comment server 1 refers to the article DB 134 and the comment DB 135 to create a data set (training data) (step S61). Each data in the data set consists of system, user, and assistant. For example, in the system, set "As an expert in biodiversity, you can make comments." In the user, set the combination of "Please create a comment as an expert in biodiversity about the following article" and the article text in the article DB 134. In the assistant, set the comment text in the comment DB 135. Create such data for all articles and use it as a data set. The control unit 11 stores the data set in the auxiliary storage unit 13 or the like (step S62). The control unit 11 transmits the data set to the dialogue system 4 (step S63). The dialogue server 41 of the dialogue system 4 receives the data set (step S64). The dialogue server 41 uses the data set to train the LLM 42 (step S65). The dialogue server 41 transmits the completion to the comment server 1 (step S66). The control unit 11 of the comment server 1 receives the completion (step S67) and ends the process. By the above process, some of the parameters of the LLM 42 that have already been optimized to a certain extent are updated, so that new knowledge is learned.

[0047] FIG. 15 is a flowchart showing another example of the generation process. A person in charge of a company inputs a business plan including the content of the business and the location where the business is to be carried out into the company terminal 2. The control unit 21 of the company terminal 2 receives the business plan (step S81). The control unit 21 transmits the received business plan to the comment server 1 (step S82). The control unit 11 of the comment server 1 receives the business plan (step S83). The control unit 11 creates a query (step S84). For example, data combining a question (inquiry) such as "Regarding the following business plan, please give comments on the impact on biodiversity." and the received business plan is created. The control unit 11 sets a role (step S85). For example, it is set as "You are an expert on biodiversity. You can answer questions such as the impact of social activities on biodiversity." The control unit 11 transmits the created role and query to the dialogue system 4 (step S86). The dialogue server 41 of the dialogue system 4 receives the role and query. The dialogue server 41 provides the role and query to the LLM 42. The LLM 42 outputs comments. The dialogue server 41 transmits the comments to the comment server 1. The control unit 11 of the comment server 1 receives the comments (step S87). The control unit 11 creates a screen including the comments (step S88). The control unit 11 transmits the screen to the company terminal 2 (step S89). The control unit 21 of the company terminal 2 receives the screen (step S90). The control unit 21 displays the screen (step S91) and ends the process.

[0048] FIG. 16 is an explanatory diagram showing an example of an article list screen. The article list screen d01 is a screen that displays a list of collected articles. The article list screen d01 includes a hot news list d011, a hot comment count d012, a new news list d013, a comment count d014, comments d015, a positive evaluation count d016, and a handle name d017. The hot news list d011 displays the titles of the top several articles in terms of the number of comments. The hot comment count d012 displays the number of comments for each article displayed in the hot news list d011. The new news list d013 displays the headings of several newly collected articles. The comment count d014 displays the number of comments for each article displayed in the new news list d013. The comments d015 display the comments of the selected article. The comments d015 are switched between display and non-display by an accordion UI (User Interface). In the example of FIG. 16, when one of the comment counts d014 is selected and clicked with a mouse or the like, the comments d015 of the corresponding article are displayed. When the corresponding comment count d014 is selected and clicked with a mouse or the like while the comments d015 are displayed, the displayed comments d015 become non-displayed. The positive evaluation count d016 indicates the number of positive evaluations called so-called "likes". When the icon to the left of the number is operated by clicking with a mouse or the like, it becomes an input of "like" for the comment. The handle name d017 indicates the handle name of the academy that entered the comment. The handle name may be entered each time a comment is entered or stored in the academy DB132 in advance.

[0049] The comment count for each article displayed on the article list screen d01 may be calculated each time by referring to the comment DB135, but it is not limited to this. A column for storing the comment count may be provided in the article DB134, and each time a comment is made, the count value of the comment may be stored in the column. The positive evaluation count for each comment may be stored in the comment DB135.

[0050] FIG. 17 is an explanatory diagram showing an example of a comment input screen. The comment input screen d02 is a screen for an academy to input comments on an article. On the article list screen d01, when the title of an article is selected by mouse click or the like, a comment input screen d02 for inputting comments on the selected article is displayed. The comment input screen d02 includes a new news list d021, a hot news list d022, an article content d023, a comment field d024, a registration button d025, and a handle name d026. The new news list d021 displays the titles of several newly collected articles. The hot news list d022 displays the titles of the top several articles in terms of the number of comments. The article content d023 displays the content of the selected article. The comment field d024 is a field for the academy to input comments. When the registration button d025 is selected, the comments input in the comment field d024 are transmitted from the academy terminal 3 to the comment server 1 and stored in the comment DB 135. The handle name d026 is the handle name of the academy. Here, it is displayed on the premise that it is preset.

[0051] The article list screen d01 shown in FIG. 16 may be used as the home screen displayed after the academy logs in to the comment system 100. The comment input screen d02 shown in FIG. 17 is an example of the input screen displayed on the academy terminal 3 in step S16 of the generation process shown in FIG. 12.

[0052] FIG. 18 is an explanatory diagram showing an example of a question-and-answer screen. The question-and-answer screen d03 is a screen where a company inputs an inquiry regarding its business and receives an answer from the comment system 100. The question-and-answer screen d03 is assumed to be used in the comment generation process shown in FIG. 13. The question-and-answer screen d03 includes a question input field d031, an execution button d032, option settings d033, a question d034, a comment d035, an article list d036, a comment d037, a number of favorable evaluations d038, and a handle name d039. The question input field d031 is a field for inputting, as a question, the content for which a comment is sought. When the execution button d032 is selected, the question input in the question input field d031 is transmitted from the enterprise terminal 2 to the comment server 1. The option settings d033 set the extraction conditions when the comment server 1 extracts comments. The said extraction conditions are used in step S45 of the comment generation process shown in FIG. 13. That is, the articles associated with the comments that satisfy the extraction conditions become the extraction targets. The question d034 shows the question transmitted to the comment server 1. The comment d035 shows the comment transmitted from the comment server 1 as an answer to the question. The article list d036 shows a list of articles extracted as articles similar to the question. The comment d037 displays the comment on the article selected in the article list d036. Similar to FIG. 16, the article list d036 is switched between display and non-display by an accordion UI (User Interface). The comment on the extracted article is transmitted from the comment server 1 to the dialogue system 4, and the summary of the comment returned from the dialogue system 4 is displayed as the above-mentioned comment d035. Regarding the comment transmitted to the dialogue system 4, only the comments that satisfy the conditions set in the option settings d033 are used. The number of favorable evaluations d038 shows the number of favorable evaluations for the comment. The handle name d039 shows the handle name of the academy that input the comment. It is also possible to let the evaluation of the comment d035 displayed on the question-and-answer screen d03 be input, for example, in three levels. The evaluation may be reflected in the evaluation of the comment d037 that refers to the said evaluation.

[0053] FIG. 19 is an explanatory diagram showing an example of a chatbot screen. The chatbot screen d04 is a screen provided by the comment server 1. The comment server 1 collaborates with the dialogue system 4 to provide a chatbot, for example, by using an API (Application Programming Interface) provided by the dialogue system 4. When providing the chatbot, it is assumed that the LLM 42 of the dialogue system 4 has been fine-tuned by the fine-tuning execution process shown in FIG. 14. The chatbot screen d04 includes a question input field d041, an execution button d042, a question d043, an answer d044, and an end button d045. The question input field d041 is a field for inputting the content for which a comment is sought as a question. When the execution button d042 is selected, the question entered in the question input field d041 is transmitted from the enterprise terminal 2 to the comment server 1. The comment server 1 transmits the question and the role to the dialogue system 4. The role does not need to be transmitted when transmitting subsequent questions. The question d043 indicates the question transmitted to the comment server 1. The dialogue server 41 of the dialogue system 4 uses the LLM 42 to generate an answer to the question and transmits it to the comment server 1. The comment server 1 transmits the answer to the enterprise terminal 2. The answer is displayed on the enterprise terminal 2. The answer is the answer d044. When the end button d045 is selected, the chatbot screen d04 is closed and the chat ends.

[0054] This embodiment has the following effects. In this embodiment, newspaper articles and magazine articles related to the collected biodiversity are presented to experts (academia) related to biodiversity, and comments are collected. The set of combinations of articles and comments becomes the knowledge of the comment system 100. The comment system 100 can use this knowledge to generate a plausible comment that would be answered by academia for an inquiry. Therefore, when a company or the like is considering a new business, it becomes possible to judge the degree of impact on biodiversity. A company or the like can refer to the comments obtained from the comment system 100 and make decisions such as continuing the business plan, revising the plan by changing the implementation location, or freezing the business plan. In addition, in order to confirm the validity of the content of the comment, it is also possible to seek advice from academia as needed. Note that the content of the business for which the inquiry is made is not limited to new ones, and may also be a business that a company is already conducting.

[0055] It is expected that, through the comment system 100, companies and the like with little knowledge about how social activities such as businesses affect biodiversity will communicate indirectly and deepen their communication with academia, which conducts daily research on the impact of social activities on biodiversity. As a result, it is considered that direct communication between companies and the like and academia will be appropriately carried out.

[0056] (Evaluation value of the comment) For the comment system 100 to create appropriate comments, the selection of comments input by academia, which is the basis, is the key. Therefore, based on the evaluation value for each comment, a selection of which comments to use is made. Specifically, when extracting comments in step S45 of the comment generation process shown in FIG. 13, those with an evaluation value smaller than a predetermined threshold are excluded. Similarly, when creating a data set in step S61 of the fine-tuning execution process shown in FIG. 14, comments with an evaluation value smaller than a predetermined threshold are not included in the data set.

[0057] Various evaluation values can be considered, and examples are shown below. The number of favorable evaluations described above, that is, the so-called "likes" number, may be used as the evaluation value as it is. Or, when the number of academics participating in the comment system 100 is set to 100, the ratio of the "likes" number can be calculated, and classification can be performed based on the obtained value, which may also be used as the evaluation value for each comment. In addition to performing the evaluation by giving "likes", multiple-level evaluations may be performed for each comment. For example, in the case of three-level evaluation, one of "good", "ordinary", or "bad" may be given.

[0058] Also, for comments with a larger number of additional comments, the evaluation value may be relatively increased or relatively decreased. If it is considered that a larger number of additional comments indicates an important comment worthy of attention, the evaluation value is increased. Otherwise, if it is considered to be a so-called "going viral" comment and adopting it will result in a lower comment accuracy, the evaluation value is decreased. Since the computer cannot determine which of the two it is, such comments may be excluded from the extraction target from the beginning, or the operator or the like may make a judgment.

[0059] In addition, the evaluation of the academy is performed, and the evaluation value of the comments created by the academy with a higher evaluation is relatively increased, and the evaluation value of the comments created by the academy with a lower evaluation is relatively decreased. The evaluation of the academy is performed, for example, by dividing the total number of "likes" collected from past comments by the number of comments to obtain the average number of "likes" per comment.

[0060] The evaluation value of the comment may be used when displaying the comment. In FIGS. 16 and 18, when displaying the comments on the selected article, the comments with higher evaluation values are displayed at the top.

[0061] (Statistical value display function) A function for presenting statistical values related to comments created by the comment system 100 to end users will be described. The statistical values to be presented are, for example, the total number of articles stored in the article DB 134, the number of articles sent to the dialogue system 4 in the generation process shown in FIG. 13, the total number of comments given to the articles, and the average number of comments per article. In the fine-tuning execution process shown in FIG. 14, the number of articles and the total number of comments included in the dataset also apply. The number of articles can be considered as a real number or a total number. When four comments are attached to one article, the total number is 4 and the real number is 1.

[0062] Statistical values regarding the academia participating in the comment system 100 can also be display targets. They are the number of people in the participating academia, the number and ratio for each specialized field. It is conceivable to display the ratio for each specialized field of the academia that has commented, the ratio for each specialized field of the academia corresponding to the comments associated with the articles sent to the dialogue system 4, and the ratio for each specialized field of the academia regarding the comments included in the dataset used for fine-tuning.

[0063] FIG. 20 is a flowchart showing an example of the procedure for field-by-field aggregation processing. The field-by-field aggregation processing is a process of obtaining the ratio for each specialized field of academia that has commented on a single article. The control unit 11 of the comment server 1 resets the counter for aggregation to 0 (step S101). The control unit 11 acquires the article ID to be processed (step S102). The control unit 11 searches the comment DB 135 and acquires the comments associated with the article ID (step S103). The control unit 11 selects one of the comments to be processed (step S104). The control unit 11 acquires the AID attached to the academia that created the selected comment (step S105). The control unit 11 stores the acquired AID in a temporary storage area provided in the main memory unit 12 or the auxiliary storage unit 13, etc. (step S106). The control unit 11 determines whether there are unprocessed comments (step S107). If the control unit 11 determines that there are unprocessed comments (YES in step S107), the process returns to step S104 to perform processing on the unprocessed comments. If the control unit 11 determines that there are no unprocessed comments (NO in step S107), it deletes the duplicates among the AIDs stored in the temporary storage area (step S108). The control unit 11 selects one AID (step S109). The control unit 11 refers to the academia DB 132 and acquires the specialized field of the academia to which the selected AID is assigned (step S110). The control unit 11 counts the counter corresponding to the acquired specialized field (step S111). If there are multiple specialized fields, each counter is counted. The control unit 11 determines whether there are unprocessed AIDs (step S112). If the control unit 11 determines that there are unprocessed AIDs (YES in step S112), the process returns to step S109 to perform processing on the unprocessed AIDs. If the control unit 11 determines that there are no unprocessed AIDs (NO in step S112), it outputs the number of people in the academia for each field and the ratio of the number of people, and ends the process (step S113). The number of people in the academia for each field and the ratio of the number of people may be updated each time a comment is given to each article and stored in the article DB 134.

[0064] FIG. 21 is an explanatory diagram showing another example of the question-and-answer screen. The question-and-answer screen d03 shown in FIG. 21 is a modified example of the question-and-answer screen d03 shown in FIG. 18. In FIG. 21, for the same configuration as in FIG. 18, the same reference numerals are given and the detailed description is omitted. In the question-and-answer screen d03 shown in FIG. 21, the number of articles d03A, the academic information button d03B, and the statistical information link d03C are added. The number of articles d03A indicates the number of articles referred to when the comment d035 is created. When the academic information button d03B is selected, information on the academies participating in the comment system 100 is displayed. When the statistical information link d03C is selected, information on the academies that made comments is displayed.

[0065] FIG. 22 is an explanatory diagram showing an example of a configuration graph of specialized fields. In FIG. 21, it is an example of a screen displayed when the academic information button d03B is selected. The graph shown in FIG. 22 shows the configuration of the specialized fields of the academies participating in the comment system 100. In the example of FIG. 22, it is shown that the academies specializing in biology are the most numerous, accounting for 30% of the total. Subsequently, the academies specializing in forest science are 20%, and the academies specializing in science are 15%. Instead of the composition ratio, the number of people may be displayed. When calculating the composition ratio, it is calculated by limiting it to the academies that created comments on the article or the academies that evaluated the comments created by other academies. It is intended not to count researchers who are not involved in the learning of the learning model M. This is because it is considered that the reliability of the comment system 100 can be obtained by showing how many experts are involved in growing the functions provided by the comment system 100, rather than simply the number of experts registered in the comment system 100.

[0066] FIG. 23 is an explanatory diagram showing an example of a specialized field composition table. In FIG. 21, it is an example of a screen displayed when the statistical information link d03C is selected. The specialized field composition table shows the composition of the academic specialized fields that commented on the article associated with the statistical information link d03C. In the example of FIG. 23, since the number of comments is 3, it is shown that one academic person each with ecology, fisheries science, and environmental science as their specialized fields has commented.

[0067] By the person in charge of the company referring to the statistical values described above, it becomes possible to numerically estimate the reliability of the comment system 100.

[0068] In the above description, it is assumed that the comment server 1 uses the external dialogue system 4. Here, the external means that the operator of the dialogue system 4 is a separate organization that has no capital relationship or cooperation relationship with the operator of the comment server 1. The operator of the comment server 1 may operate the dialogue system. For example, the learning model M of the comment server 1 is a large language model, and the comment server 1 may provide the dialogue system 4 using the learning model M.

[0069] The technical features (constituent elements) described in each embodiment can be combined with each other, and by combining them, new technical features can be formed. The embodiments disclosed this time should be considered as illustrative in all respects and not restrictive. The scope of the present invention is shown not by the above meaning, but by the claims, and all changes within the meaning and scope equivalent to the claims are intended to be included. Also, although the claims use a format (multi-claim format) in which claims that cite two or more other claims are described, it is not limited to this. It may be described using a format in which a multi-claim (multi-multi-claim) that cites at least one multi-claim is described.

Explanation of Reference Numerals

[0070] 100: Comment System 1: Comment Server 11: Control Unit 12: Main Memory Unit 13: Auxiliary Memory Unit 131: Enterprise DB 132: Academia DB 133: Thesaurus DB 134: Article DB 135: Comment DB 136: Knowledge DB 14: Communication Unit 15: Reading Unit 1P: Control Program 1a: Portable Memory Medium 1b: Semiconductor Memory 2: Enterprise Terminal 21: Control Unit 22: Main Memory Unit 23: Auxiliary Memory Unit 24: Communication Unit 25: Input Unit 26: Display Unit 2P: Control Program 3: Academia Terminal 31: Control Unit 32: Main Memory Unit 33: Auxiliary Memory Unit 34: Communication Unit 35: Input Unit 36: Display Unit 4: Dialogue System 41: Dialogue Server 42: LLM B: Bus N: Network

Claims

1. Obtain inquiries including activity details, Comments on biodiversity in response to the acquired inquiries are obtained using a learning model, Output the retrieved comments An information processing method in which processing is performed by a computer.

2. The learning model is trained using training data in which article information including the content of an activity is associated with comments from pre-designated experts about biodiversity for the activity based on the article information. The information processing method according to claim 1 .

3. The training data includes ratings for the comments. The information processing method according to claim 2 .

4. the learning model includes a language model; The language model outputs a comment for the acquired query based on a database that associates the acquired query with article information including the content of the activity and comments on biodiversity from pre-specified experts. The information processing method according to any one of claims 1 to 3.

5. Obtaining a query that includes a description of the activity and a location of the activity. The information processing method according to any one of claims 1 to 3.

6. Output the number of article information used for learning, the actual number or percentage of experts who evaluated the comments or experts who created the comments by field. The information processing method according to any one of claims 1 to 3.

7. Obtain inquiries including activity details, Comments on biodiversity in response to the acquired inquiries are obtained using a learning model, Output the retrieved comments An information processing program that causes a computer to execute a process.

8. An information processing device including a control unit, The control unit Obtain inquiries including activity details, Comments on biodiversity in response to the acquired inquiries are obtained using a learning model, Output the retrieved comments An information processing device that executes processing.

Images