Chatbot agent service enabling question and answer related to specific law and regulation

Abstract

To allow a user to quickly and appropriately obtain calculations and other procedures and information related to procedures for applications specified in an act on the improvement of energy consumption performance of buildings, an enforcement order of the act on the improvement of energy consumption performance of buildings, and ministerial ordinance that specifies energy consumption performance standards for buildings (hereinafter referred to as Act on the Improvement of Energy Consumption Performance of Buildings, and the like).SOLUTION: A system is configured to allow a chatbot integrated with artificial intelligence (large language model) to quickly and appropriately give an answer, through a web application from a user terminal, to a question from a user about information related to calculations and other procedures and procedures for applications specified in Act on the Improvement of Energy Consumption Performance of Buildings, and the like.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention is a chatbot agent service that can answer questions with high accuracy regarding specific laws and regulations, consisting of an artificial intelligence with closed re-training authority, a chatbot, and an application installed on a user's terminal.The specific laws and regulations are the Act on Improvement of Energy Consumption Performance of Buildings, the Enforcement Order of the Act on Improvement of Energy Consumption Performance of Buildings, and the Ministerial Ordinance that Establishes Building Energy Consumption Performance Standards, making it a highly specialized chatbot agent service. [Background technology]

[0002] A large-scale language model is a machine learning model trained on large amounts of text data to mimic human language comprehension. It is trained on large text datasets containing billions or more words and sentences, including web text, books, articles, and conversation logs. It is capable of making complex decisions, such as correlating learned knowledge, and is also capable of understanding and generating text data in natural language. The text data generation process takes into account understanding the meaning of the text, consideration of context, and adherence to grammar, making it a promising technology for applications such as text summarization, text translation, and question-and-answering. Another feature of the learning process is that it has an open domain where it can be retrained with input from an unspecified number of users, and a closed domain where it cannot be retrained with input from an unspecified number of users. Chatbots are now commonplace thanks to the development of internet infrastructure and the evolution of user devices. Some simply respond to specific input strings of characters regardless of context and return specific sentences (such as spam on social media), while others are designed by chatbot service providers to limit and guide the content of Q&A sessions to suit the specific services they offer (such as a delivery company's redelivery procedure service). Recently, chatbots have also emerged that combine the above-mentioned large-scale language models to allow for complex conversations (AI chat). The Act on the Improvement of Energy Consumption Performance of Buildings, the Enforcement Order of the Act on the Improvement of Energy Consumption Performance of Buildings, and the Ministerial Ordinance Prescribing Building Energy Consumption Performance Standards (hereinafter referred to as the Building Energy Conservation Act, etc.) can be said to be legal systems established to enable the construction sector to fulfill the role assigned to it towards achieving carbon neutrality by 2050. When constructing a new building, thresholds are set for the building envelope performance and primary energy consumption, etc., and these laws are so binding that buildings cannot be constructed unless they demonstrate performance above these thresholds. [Prior art documents] [Patent documents]

[0003] Nothing in particular [Non-patent literature]

[0004] Nothing in particular Summary of the Invention [Problem to be solved by the invention]

[0005] The Japanese government has set a goal of achieving carbon neutrality by 2050. Estimates using the 1990 industry-related table and other data suggest that construction-related emissions account for 36.0% of Japan's CO2 emissions, indicating that the construction industry has a major role to play in achieving carbon neutrality by 2050. Against this background, regulations are being rapidly strengthened, with the Energy Conservation Law for Buildings being separated from the Energy Conservation Law in 2016, which applies only to buildings, and full compliance with energy conservation standards being made mandatory by 2025. On the other hand, architects, construction companies, and other construction-related businesses, both large and small, are concerned about the increased burden of work related to the Building Energy Conservation Act in addition to the existing procedures. Furthermore, an increase in problems, such as a lack of communication and misunderstandings with building owners, is also expected regarding matters related to the Building Energy Conservation Act. [Means for solving the problem]

[0006] One way to solve these problems is to have a chatbot agent service that specializes in the Building Energy Conservation Act, etc. The most important point of this patent application, and the novelty of the invention, is that among the wide range of building code-related regulations, including the Building Standards Act, we chose to specialize in the Building Energy Conservation Act, etc. In an experiment organized by an industry newspaper, ChatGPT took the first-class architect's exam. The correct answer rate for the legal section of the planning, facilities, regulations, structure, and execution sections was about 50%. While the test questions may differ slightly from natural language, this level of accuracy is not surprising when asked to solve a wide range of complex problems. Furthermore, actual laws and regulations are often detailed and set by specific local administrative agencies in the form of ordinances and collections of regulations. Taking these factors into consideration, it is clear that providing an agent service for the Building Standards Act and related regulations (including the Building Energy Conservation Act, etc.) in general is quite difficult. However, if we limit it to the Building Energy Conservation Act, etc., the content is limited, and since the handling does not differ between specific administrative agencies in each region, a fairly uniform response is possible, making it a realistic scope for providing agent services. Such judgments are limited to people who have sufficient knowledge of the Building Standards Act and related regulations (including the Building Energy Conservation Act, etc.) in general and review practices, as well as a certain level of knowledge about application services, an interest in cutting-edge fields such as ChatGPT, and who are in an environment where they can actively create services, such as being the representative of a corporation, and who can make a comprehensive judgment and control the system construction toward service creation. This can be said to be an idea with a high level of expertise, specificity, novelty, and advanced thinking. [Effects of the Invention]

[0007] If there were a system that could provide accurate answers to questions about application procedures, calculation procedures, etc., it would be possible to dramatically reduce the labor costs of construction-related contractors. Furthermore, by sharing the chatbot agent service with building owners, it would be possible to reduce communication gaps and discrepancies between building owners and construction-related contractors, leading to reduced communication costs. These efforts would reduce the hurdles to making buildings more energy-efficient, and could become the infrastructure that allows building owners and construction-related contractors to work together toward carbon neutrality by 2050. [Brief explanation of the drawings]

[0008] [Figure 1] Figure 1 is an explanatory diagram showing an overview of a chatbot agent service specialized in the Building Energy Conservation Act and other laws. DETAILED DESCRIPTION OF THE INVENTION

[0009] Below, with reference to the accompanying drawings, an embodiment of a chatbot agent service specialized for the Building Energy Conservation Act and the like will be described in detail. [Industrial Applicability]

[0010] As mentioned above. [Explanation of symbols]

[0011] 1 user 2. User terminal 3. Applications 4. Chatbots 5. Artificial Intelligence 6. Artificial Intelligence Educator

Claims

1. A customer support system that consists of artificial intelligence using a large-scale language model (excluding packages that are retrained by an unspecified number of inputs), a chatbot, and an application installed on a user's device such as a smartphone or PC. When a user inputs a question to a chatbot using an application installed on the user's device, the chatbot transfers the question to an API-linked artificial intelligence, the artificial intelligence trained by prompt engineering thinks of an answer and returns the answer to the chatbot, and the chatbot displays the answer returned by the artificial intelligence using an application installed on the user's device, thereby providing customer support. The system is limited to answering questions related to calculation procedures and application procedures stipulated in the Act on Improvement of Energy Consumption Performance of Buildings, the Enforcement Order of the Act on Improvement of Energy Consumption Performance of Buildings, and the Ministerial Ordinance Establishing Energy Consumption Performance Standards for Buildings, etc.

2. The artificial intelligence uses a package that cannot be retrained by input from an unspecified number of people, and therefore training is left to the discretion of the service provider.In the prompt engineering for training, the training was limited to those related to calculation procedures, application procedures, etc. stipulated in the Act on Improvement of Energy Consumption Performance of Buildings, the Enforcement Order of the Act on Improvement of Energy Consumption Performance of Buildings, and the Ministerial Ordinance Establishing Energy Consumption Performance Standards for Buildings, etc.

Images