Farm response system and farm response program

Abstract

This provides farming response technology that can appropriately answer various questions tailored to the individual circumstances of farmers. [Solution] The farming response system includes a question processing unit 20 that acquires user farming data and general farming data related to the question based on information received from the user regarding questions about farming; an answer request creation unit 31 that creates an answer request to obtain an answer from an answer calculation unit 70 based on the question, user farming data, and general farming data; and an answer notification unit 32 that provides the user with answer information generated based on the answer from the answer calculation unit 70.

Description

Technical Field

[0001] The present invention relates to a farming response system and a farming response program that receive farming-related questions from users and return answers to those questions to the users.

Background Art

[0002] As a system for answering (proposing) questions (doubts) of users that occur when operating agriculture, Patent Document 1 discloses an agricultural support device that proposes farming operations for producing agricultural crops in a field based on a production plan and proposes promotion or suppression of the demand for agricultural crops based on the production plan.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In the agricultural support device according to Patent Document 1, the demand proposal unit proposes promotion or suppression of the demand for agricultural crops based on the production plan, and this proposal is output (transmitted) to the consumer terminal by the communication unit. At that time, the proposal for promoting the demand for agricultural crops includes information (such as a message) indicating promotion of the sale of agricultural crops. The proposal for suppressing the demand for agricultural crops includes information (such as a message) indicating suppression of the sale of agricultural crops. Further, the production proposal unit proposes the work or agricultural machinery necessary for producing agricultural crops in the field, and also proposes the number of workers for producing agricultural crops in the field.

[0005] The agricultural support device disclosed in Patent Document 1 is beneficial to farmers because it predicts crop demand and the resources necessary for farming based on production plans derived from historical data on agricultural work, through statistical calculations. However, such an agricultural support device cannot answer personal questions for farmers, such as "I want to know what I did yesterday" or "When and where should I refuel my farm machinery?", nor can it answer questions that cannot be derived simply from farming performance data, such as "Are there any subsidies I can use, and how should I apply for them?" or "Where can I get advice about the next planting plan?".

[0006] In light of these circumstances, the object of the present invention is to provide farming response technology that can appropriately answer various questions tailored to the individual circumstances of farmers. [Means for solving the problem]

[0007] The farming response system according to the present invention comprises: a question processing unit that receives a question from a user regarding farming and acquires the user's farming data and general farming data related to the question based on the received question information; a response request creation unit that creates a response request to obtain an answer to the question from a response calculation unit based on the question, the user's farming data, and the general farming data; and a response notification unit that provides the user with response information generated based on the answer from the response calculation unit.

[0008] In this configuration, user farming data, which is the personal farming information of the user who asked the question, and general farming data related to the words included in the question are acquired. Based on this user farming data and general farming data, an answer request (e.g., a prompt) is created that asks the answer calculation unit to answer the question. Since the answer request includes user farming data and general farming data, the answer calculation unit is able to output an answer that satisfies the user who asked the question. The obtained answer is given to the user as answer information, either in its original state, or after some transformation, or with the addition of illustrations or tables. If the user's question is only about personal farming information, such as "I want to know what I did yesterday," the general farming data may be omitted, and the answer request may be created using only user farming data. The answer calculation unit here can preferably be configured as a generating AI, and particularly preferably as a combination of LLM (Large Language Model) and RAG (Retrieval-Augmented Generation).

[0009] Furthermore, depending on the question, it may be possible to answer by referring only to the user's farming data. Therefore, a farming response system that answers by referring only to the user's farming data is also a useful farming response system. Accordingly, the present invention also proposes a farming response system comprising: a question processing unit that receives a question about farming from a user and acquires the user's farming data, or the user's farming data and general farming data related to the question, based on the information received from the question; an answer request creation unit that creates an answer request to obtain an answer to the question from an answer calculation unit based on the question and the user's farming data, or the question, the user's farming data and the general farming data; and an answer notification unit that provides the user with answer information generated based on the answer from the answer calculation unit.

[0010] The individual information of the user who asks the question includes the user's address and name, the farmland and farm machinery owned by the user, work history information showing the farm work the user has performed, and the farm work plan the user will perform in the future. In addition, general farming materials related to the content of the question are materials necessary to answer a wide range of questions, and include, for example, laws and regulations concerning farming managed by the national and local governments, recommended farming items, farming subsidies, as well as rules and recommended farming items managed by organizations such as agricultural cooperatives, and papers managed by universities and academic societies. Thus, in this invention, it is proposed that the user farming materials include at least one of the user's address and name information, farm machinery ownership information, farmland ownership information, farming history information, and farming plan information, and that the general farming materials include at least one of the farming work rules, farming subsidy grant rules, farming recommendation events, and farming papers from agricultural government agencies and agricultural organizations.

[0011] In a system such as the farming response system according to the present invention, users are generally registered, and users can ask questions by logging in as a registered user. Therefore, users can be identified by the user code (identification information that identifies the user) issued to the registered user. Once a user is identified, the recorded information of that user (individual user information) linked to that user code can be extracted using the user code and other user search conditions. Thus, in the present invention, it is proposed that the question processing unit extracts the user's farming data from the user management database using the user code that identifies the user and the user search conditions derived from the question.

[0012] Since the source data for general farming materials is enormous, it is important to accurately and quickly extract only the farming materials relevant to the question. For this purpose, it is preferable that appropriate materials be extracted from various databases as search conditions created from words, phrases, and strings included in the question. Therefore, in this invention, it is proposed that the question processing unit extracts the general farming materials from the agricultural information database using general search conditions derived from the question.

[0013] When extracting farming data related to a question, if only the words and strings contained in the question are used as search criteria, data containing similar words may not be extracted. For this reason, it is preferable to include similar words and strings contained in the question as search criteria. A suitable example of creating search criteria that include similar words is to vectorize the words and strings contained in the question, and then derive search criteria that include similar words by performing a vector operation (such as a dot product) between the vectorized question and a pre-vectorized set of search criteria. For this reason, the present invention proposes that the general search criteria are derived by performing a vector operation between the vectorized question and a pre-set set of vectorized general search criteria.

[0014] In some cases, it may be preferable for multiple different answers to be output for each question. Therefore, this invention proposes that the answer may contain multiple answer elements. This allows the user to interpret the answer from multiple perspectives.

[0015] When multiple answer elements are available, and the user desires a quick and simple answer, it is preferable to select and provide the user with the most appropriate answer element. A suitable example of such a selection method is to select the answer element based on the user's unique information, such as whether they are a large-scale or small-scale farmer, whether they cultivate multiple or few varieties of crops, or whether they are a rural or suburban farmer. Therefore, in this invention, it is also proposed that the answer notification unit generate the answer information from at least one of the answer elements that are determined to be appropriate based on the user's farming data.

[0016] This invention applies not only to farming response systems but also to computer programs used to construct farming response systems. Such a farming response program causes a computer to execute the following functions: a question processing function that receives farming-related questions from a user and acquires the user's farming data or general farming data related to the question, or both; an answer request creation function that creates an answer request to obtain an answer from an answer calculation unit based on the question and the user's farming data or the general farming data, or both; and an answer notification function that provides the user with answer information generated based on the answer from the answer calculation unit. Such a farming response program also has the various functions and effects of the farming response system described above. [Brief explanation of the drawing]

[0017] [Figure 1] This is a schematic diagram of the agricultural response system. [Figure 2] This is a functional block diagram of the farming response system. [Figure 3] This is a flowchart illustrating an example of data flow in a farm management response system. [Modes for carrying out the invention]

[0018] The schematic configuration of the farming response system according to the present invention will be explained with reference to Figure 1.

[0019] The core component of the farming response system is the user interface unit 2, which is configured as a computer system or a functional unit built into a computer system. Peripheral systems that work in conjunction with this user interface unit 2 include the user terminal 1, user services 5, multiple agricultural information databases 6, and cloud services 7 that function as generating AI services. The user terminal 1, user interface unit 2, user services 5, and cloud services 7 are connected to each other via LAN lines, data communication lines, the internet, etc., enabling data exchange.

[0020] The user terminal 1 is a smartphone, a tablet computer, a personal computer, etc., creates a question that the user wants to ask, sends it to the user interface unit 2, receives the answer information sent from the user interface unit 2, and notifies a monitor or the like.

[0021] The core element of the user service 5 is, in this embodiment, a service computer system that manages the farming information of registered users. This farming information includes various information related to the user's farming, such as user-specific information linked to a user code as identification information for identifying the user, for example, the user's address and name information, the user's agricultural machinery ownership information, the user's farmland ownership information, farming history information indicating the farming operations carried out by the user, the number of workers indicating the number of people involved in the farming operations, and farming plan information that the user will carry out in the future. These information are stored in the system so that they can be extracted by the user or the person with management authority. Note that the various information related to the user's farming is not limited to these and includes all information related to farming and agricultural operations. Note that, as will be described in detail later, the farming information of these users (user farming information) is treated as user farming materials for those stored in the user management database 50.

[0022] The agricultural information database 6 is a server system constructed in public or private agricultural-related facilities scattered across various locations, and has a function of providing various general farming materials (this general farming material is a term used to distinguish from the above-mentioned user farming information, and in this application, many types of agricultural materials related to farming are handled). The general farming materials include, for example, general cropping methods and conditions of crops, general information on farming and agricultural operations, laws and farming operation rules related to farming managed by the country, agricultural authorities, and local governments, or farming recommendation events and items, rules for stipulating subsidies related to farming such as farming subsidy delivery rules, and further include rules and recommended farming items managed by agricultural organizations such as agricultural cooperatives, and farming papers managed by universities and academic societies. These general farming materials include those in which public or private agricultural-related facilities, etc. independently collect and manage various rules and items. In FIG. 1, a plurality of agricultural information databases 6 are shown. Note that a dedicated database 60 newly constructed by importing specific general farming materials required from such a plurality of agricultural information databases 6 can replace the agricultural information database 6.

[0023] The cloud service 7 is constructed by a computer system, and a generative AI service (generative AI) as a response calculation unit 70 is constructed within the computer system. In this embodiment, the generative AI is an LLM which is a large language model and is capable of receiving RAG. The cloud service 7, that is, the response calculation unit 70, responds to a response request from the user interface unit 2 and returns the calculated response to the user interface unit 2.

[0024] The user interface unit 2, which will be described in detail later, receives a farming-related question from the user terminal 1, obtains the user's user farming materials and general farming materials based on the received farming-related question, sends a response request created based on the question, the user farming materials, and the general farming materials to the response calculation unit 70, and directly or after performing some processing on the response obtained from the response calculation unit 70, transmits it as response information to the user terminal 1.

[0025] The schematic configuration diagram in Figure 1 is an example; for instance, the user interface unit 2 can be built within user service 5 as one of the functional components of the computer system that constitutes user service 5. Similarly, a dedicated database 60 that replaces multiple agricultural information databases 6 can also be built within user service 5.

[0026] Next, we will explain the functional parts of the user interface unit 2 using the functional block diagram in Figure 2.

[0027] The user interface unit 2 comprises a question processing unit 20, an answer request creation unit 31, and an answer notification unit 32.

[0028] The question processing unit 20 has a user management unit 21 and a farming data management unit 22 to process questions sent from the user terminal 1. When the user management unit 21 receives a question from the user terminal 1, it obtains the user's login information, which is the reception information (which includes user identification information such as a user code), during the login process, and creates search conditions (which are the user search conditions in this invention) based on this user code and the content of the question. For example, these search conditions can be created from the user code and the words of the question. As an example, if the question is "I want to know what I did yesterday," the search conditions are created from the user code and a keyword that can specify the latest work history file linked to this user code (it may simply be "work content"). The user management unit 21 provides a query using these search conditions to the user management database 50 built in the user service 5. In response to this query, the user management database 50 extracts the contents (at least the latest contents) of the work history file that records the various farming activities of the user, and sends them to the user management unit 21 as user farming data.

[0029] The Agricultural Data Management Unit 22 creates search conditions (general search conditions in this invention) based on the user's farming area and the content of the question. The search conditions here are the words, strings, and similar words used in the question, but if there is a text interpretation function, the result of interpreting the question may also be used. If a lookup table is provided for deriving general search conditions from the farming area and the content of the question, the general search conditions may be derived from the user's farming area and the content of the question using such a lookup table. Instead of a lookup table, the search conditions may be derived using a vector operation (dot product) between a vectorized question and a group of vectorized general search conditions. In addition, multiple items necessary for deriving general search conditions may be combined. In any case, the Agricultural Data Management Unit 22 provides the agricultural information database 6 with a query using the derived search conditions. The agricultural information database 6 responds to this query by extracting general agricultural data that matches the search conditions and sending it to the Agricultural Data Management Unit 22. General farming resources vary depending on the user's question, but they may include laws and regulations related to farming managed by the national and local governments, recommended farming items, farming subsidies, as well as regulations and recommended farming items managed by organizations such as agricultural cooperatives, and papers managed by universities and academic societies.

[0030] The response request creation unit 31 creates a response request to obtain an answer to the user's question from the response calculation unit 70, which is an LLM (Low-Level Machine), based on the question received from the user terminal 1, the user farming data received from the user management unit 21, and the general farming data received from the farming data management unit 22.

[0031] The response request created by the response request creation unit 31 includes a prompt containing the user's questions and reference materials (user farming data and general farming data). However, if the reference materials are included in the prompt, the reference materials are omitted, and the prompt becomes the response request.

[0032] The answer calculation unit 70 of the cloud service 7 generates an answer based on the answer request sent from the answer request creation unit 31 of the user interface unit 2. The generated answer is sent to the answer notification unit 32 of the user interface unit 2.

[0033] The response notification unit 32 generates response information based on the received response and transmits it to the user terminal 1. The response information may be the response itself received from the response calculation unit 70, or it may include tables, illustrations, and other additional information generated based on the response. If the obtained response consists of multiple response elements, the response notification unit 32 may be equipped with a function to select the most appropriate response element from the question content and user attributes and provide it to the user.

[0034] Next, we will explain an example of data flow in the farming response system using the flowchart in Figure 3.

[0035] A user logs in to the user interface unit 2 via user terminal 1 and asks a question about farming (#01). The question processing unit 20 of the user interface unit 2 receives this question, creates user search conditions from the user information of the logged-in user, and accesses the user management database 50 using these user search conditions as a query (#02). The user management database 50 extracts user farming data based on this query and provides it to the user interface unit 2 (#03). Furthermore, the question processing unit 20 creates general search conditions from the content of the received question and accesses the agricultural information database 6 using these general search conditions as a query (#04). The agricultural information database 6 extracts general farming data based on this query and provides it to the user interface unit 2 (#05). Steps #02 and #04 may be performed simultaneously, or step #04 may be performed before step #02.

[0036] The answer request creation unit 31 of the user interface unit 2 combines the received question, user farming data, and general farming data to create an answer request for the answer calculation unit 70, which is an LLM, and provides it to the answer calculation unit 70 (#06). The answer calculation unit 70 outputs an answer based on the given answer request and provides it to the user interface unit 2 (#07). The answer notification unit 32 of the user interface unit 2 converts (generates) the answer received from the answer calculation unit 70 into answer information and transmits it to the user terminal 1 (#08).

[0037] In the farming response system described above, the question processing unit 20 was configured to acquire user farming data and general farming data related to the question based on the received question information, and the answer request creation unit 31 was configured to obtain an answer to the question from the answer calculation unit based on the question, user farming data, and general farming data. However, instead, the question processing unit 20 may be configured to acquire only user farming data based on the received question information, and the answer request creation unit 31 may be configured to obtain an answer to the question from the answer calculation unit based on the question and user farming data. In such a farming response system, for questions where an answer can be sufficiently obtained using only user farming data, a quick answer can be obtained by omitting the acquisition and reference of general farming data. Furthermore, the present invention does not exclude a farming response system configured in which the question processing unit 20 acquires only general farming data based on the received question, and the answer request creation unit 31 obtains an answer to the question from the answer calculation unit based on the question and general farming data. For example, such a farming response system is suitable for learning about general farming laws, regulations, and proposals.

[0038] [Another embodiment] (1) In the above-described embodiment, the answer calculation unit 70 was configured as a generative AI combining LLM and RAG, but the present invention is not limited thereto, and may be configured as a system that allows the LLM to learn unique information, such as a system that processes only prompts or a dedicated system that has been fine-tuned in advance.

[0039] (2) The functional parts of the user interface unit 2 shown in Figure 2 are primarily descriptive, and each functional part can be integrated with other functional parts, and each functional part can also be further subdivided.

[0040] (3) In the above-described embodiment, the answer calculation unit 70 was built on the computer system of the cloud service 7, but instead, the answer calculation unit 70 may be built on the computer system of the user service 5. Furthermore, the user interface unit 2, the answer calculation unit 70, and a dedicated database 60 that functions as a reduced version of the agricultural information database 6 may be built on the computer system of the user service 5.

[0041] (4) User questions may be in the form of messages or voice. In that case, the user interface unit 2 may be equipped with a voice recognition unit, a voice synthesis unit, etc.

[0042] Furthermore, the configurations disclosed in the above embodiments (including other embodiments, the same applies hereinafter) can be applied in combination with configurations disclosed in other embodiments, as long as no inconsistencies arise. Moreover, the embodiments disclosed herein are illustrative, and the embodiments of the present invention are not limited thereto, and can be modified as appropriate without departing from the object of the present invention. [Industrial applicability]

[0043] This invention is applicable to inquiry technology in the agricultural field. [Explanation of symbols]

[0044] 1: User terminal 2: User Interface Unit 5: User Services 6: Agricultural Information Database 7: Cloud Services 20: Question Processing Section 21: User Management Department 22: Agricultural Data Management Department 31:Response request creation department 32: Answer from the News Department 50: User Management Database 60: Dedicated Database 70: Answer calculation unit AI: Generation

Claims

1. A farming response system, A question processing unit receives questions from users regarding farming, and based on the information received from the questions, obtains the user's farming data and general farming data related to the questions. A response request creation unit creates a response request to obtain an answer to the question from a response calculation unit based on the aforementioned question, the user farming data, and the general farming data. A response notification unit provides the user with response information generated based on the response from the response calculation unit. A farming response system equipped with the following features.

2. A farming response system, A question processing unit receives questions from users regarding farming, and based on the information received from the questions, obtains the user's farming data, or user farming data and general farming data related to the questions. A response request creation unit creates a response request to obtain an answer to the question from a response calculation unit based on the aforementioned question and the user farming data, or the aforementioned question, the user farming data, and the general farming data. A response notification unit provides the user with response information generated based on the response from the response calculation unit. A farming response system equipped with the following features.

3. The farming response system according to claim 1 or 2, wherein the user farming data includes at least one of the user's address and name information, farming machinery ownership information, farmland ownership information, farming history information, and farming plan information, and the general farming data includes at least one of the farming work rules of agricultural government agencies and agricultural organizations, farming subsidy grant rules, farming recommendation events, and farming papers.

4. The farming response system according to claim 1 or 2, wherein the question processing unit extracts user farming data from a user management database using a user code that identifies the user and user search conditions derived from the question.

5. The farming response system according to claim 1 or 2, wherein the question processing unit extracts the general farming data from the agricultural information database using general search conditions derived from the question.

6. The farming response system according to claim 5, wherein the general search conditions are derived by a vector operation between the vectorized question and a pre-set set of vectorized general search conditions.

7. The farming response system according to claim 1 or 2, characterized in that the aforementioned response includes multiple response elements.

8. The farming response system according to claim 7, wherein the response notification unit generates the response information from at least one of the response elements that are determined to be appropriate based on the user farming data.

9. A farming response program, A question processing function that receives questions from users regarding farming and retrieves the user's farming data, general farming data related to the question, or both. A response request creation function that creates a response request to obtain an answer to the question from a response calculation unit based on the aforementioned question and the user farming data or the general farming data or both. A response notification function that provides the user with response information generated based on the response from the response calculation unit, A farming response program that causes a computer to execute commands.

Images