AI-based disaster safety knowledge integrated management system

The AI-powered disaster safety knowledge integrated management system addresses fragmented data issues by providing a question-and-answer service for expert knowledge, reducing costs and enabling efficient policy planning and reporting.

JP2024543281AActive Publication Date: 2024-11-21NAT DISASTER MANAGEMENT INST
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
JP2023550230
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Priority Date
2022-10-06
Filing Date
2023-03-09
Publication Date
2024-11-21
Estimated Expiration
2043-03-09

AI Technical Summary

Technical Problem

Existing disaster safety management systems lack integrated data management, leading to fragmented and difficult-to-utilize information, reliance on experiential knowledge, and labor-intensive decision-making, which results in high costs and analysis blind spots.

Method used

A disaster safety knowledge integrated management system using AI that provides a question-and-answer service for expert knowledge, supporting policy planning and reporting through data collection, analysis, and automatic reporting services.

Benefits of technology

Reduces time, manpower, and cost required for data exploration and reporting, enabling non-experts to generate analysis and policy materials with machine assistance, improving disaster safety policy support through an integrated knowledge base.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2024543281000001_ABST
    Figure 2024543281000001_ABST
Patent Text Reader

Abstract

This invention relates to an AI-based disaster safety knowledge integrated management system that utilizes an intelligent analysis service of disaster safety data to enable an AI-based question-answering service of specialized knowledge in the field of disaster safety and an automatic reporting service that supports the creation of policy plans and report materials for specific subjects. It is composed of a disaster safety knowledge base part connected to a data network and an artificial intelligence part for realizing high-dimensional information processing using artificial intelligence. The disaster safety knowledge base unit is composed of a data collection unit for collecting and aggregating various information from external organizations; a data transmission unit for transmitting the aggregated information to a server via LTE, 5G, WIFI, etc.; and a big data unit for analyzing and storing the transmitted data. The artificial intelligence unit utilizes the data accumulated and analyzed in the big data unit to make judgments and inferences based on human cognitive abilities (language, voice, vision, sensitivity, etc.) and learning and inference functions, thereby featuring machine intelligence through rapid learning using data.
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present invention relates to an AI-based integrated disaster safety knowledge management system that utilizes an intelligent analysis service of disaster safety data to enable an AI-based question-answering service of specialized knowledge in the field of disaster safety and an automatic reporting service that supports policy planning and the creation of report materials for specific topics. [Background technology]

[0002] The Fourth Industrial Revolution will bring about a dramatic increase in productivity through the intelligence of machines and a fundamental transformation of the industrial structure, made possible by changes in intelligent information technology.

[0003] Intelligent information technology is a combination of "intelligence," which uses artificial intelligence to realize high-dimensional human information processing, and "information" based on data network technology (IoT, cloud, big data, mobile). Artificial intelligence is a technology that realizes human cognitive abilities (language, voice, vision, sensitivity, etc.) as well as intelligence such as learning and inference, and encompasses artificial intelligence software / hardware and basic technologies. Data network technology serves as an important foundation supporting the rapid performance improvement, spread and diffusion of artificial intelligence technology, and is an essential ICT technology for generating, collecting, transmitting, storing and analyzing data.

[0004] Intelligent information technology has developed to a level where machines have the ability to make unmanned decisions (exercising high-level human judgment abilities), react in real time, evolve autonomously, and learn to digitize everything, and is now being applied to specialized fields.

[0005] Recently, the government has been placing emphasis on utilizing intelligent information technology at a national level in order to promote the shared use of data among government agencies and public institutions, and to establish an integrated management system to stimulate objective and scientific policy-making and decision-making, and is currently developing related laws and policies.

[0006] On the other hand, database-based disaster safety management is urgently needed to efficiently respond to new types of disasters, complex disasters, and other disasters, and to formulate scientific and systematic safety policies. In order to establish a legal and institutional foundation for systematically managing data for database-based disaster management, and to aim for private sharing and opening up, it is necessary to build integrated disaster safety data.

[0007] "Integrated disaster safety data" is a database-based data that will be used to promote disaster safety policies, such as for damage prediction and the selection of vulnerable safety areas. It consists of a process for collecting and linking data, a process for accumulating and storing this data, and a process for releasing and utilizing it for those who need it.

[0008] Now that data-centric government policies and related policies are being promoted in earnest, there is an urgent need to replace traditional decision-making methods based on experience and intuition with objective, scientific, data-based administrative methods to increase credibility and improve the quality of life of the people (see Figure 9).

[0009] However, the necessary data is fragmented and difficult to use as information, and not only has empirical knowledge not been digitized, but an integrated management system has not been established, meaning that the disaster safety management database is not being used appropriately.

[0010] In addition, traditional decision-making methods based on experience and intuition require labor-intensive intellectual activities such as policy planning and decision-making, which require significant time and personnel costs. Furthermore, there are problems with relying on skilled experts and having blind spots in analysis, and there are problems with the need for advanced intellectual labor, from researching and searching materials necessary for policy reports to writing reports.

[0011] In other words, there are limitations to the direct analysis that beginners or those without experience can perform, and there is a high degree of reliance on highly trained, specialized analysts, which creates blind spots in the analysis and makes it impossible to perform an integrated analysis of all related areas. Summary of the Invention [Problem to be solved by the invention]

[0012] The first objective of the present invention is to utilize disaster safety data sharing and intelligent analysis services of disaster safety data to enable an AI-based question-answering service of specialized knowledge in the field of disaster safety, and to enable an automated reporting service that supports policy planning and the creation of report materials on specific topics.

[0013] Another objective is to dramatically reduce the time, manpower, costs and workload required for data exploration, analysis and reporting; enable non-experts or people with relatively little experience to perform analysis and generate policy materials with the assistance of machines; and help make comprehensive decisions by taking into account information other than disasters, new technologies and new issues. In this way, the level of disaster safety policy support will be significantly improved through insights gained from an integrated knowledge base in the field of disaster safety. [Means for solving the problem]

[0014] In order to achieve the above objectives, the AI-based disaster safety knowledge integrated management system of the present invention utilizes disaster safety data sharing and intelligent analysis services of disaster safety data to enable an AI-based question-answering service for specialized knowledge in the field of disaster safety, and is related to an AI-based disaster safety knowledge integrated management system that enables an automatic reporting service that supports policy planning and the creation of report materials for specific topics.

[0015] The disaster safety knowledge integrated management system consists of a disaster safety knowledge base section connected to a data network, and an artificial intelligence section that uses artificial intelligence to realize high-dimensional human information processing.

[0016] The Disaster Safety Knowledge Base Unit includes a data collection unit for collecting and aggregating various information from external agencies, a data transmission unit for transmitting the collected information to a server via LTE, 5G, WIFI, etc., and a big data unit for analyzing and storing the transmitted data.

[0017] The data collection section provides general materials such as disaster status, disaster statistics, disaster policies, status reports, and research reports, as well as unstructured data such as news, SNS, laws and regulations, status reports, and research reports, and structured data such as damage recovery, weather, safety index, disaster sales, and 119 reports. The big data section also classifies, analyzes, and accumulates data provided by the data transmission section, and classifies it into real-time data such as steep slope sensors, water level information, and photos of the disaster site, structured data such as past damage history information and safety information of various facilities, and unstructured data such as disaster status report text, disaster status report images, and briefing materials. The data classified by these methods is classified into disaster accident damage status, disaster accident response history, and disaster safety policy information, and is accumulated as archive data, so that the data can be made open and shared with necessary targets (general public, related organizations, etc.). The artificial intelligence section also utilizes the data accumulated and analyzed by the big data section to perform judgment and inference using human cognitive abilities (language, voice, vision, sensibility, etc.) and learning and inference functions, and is characterized by machine intelligence through rapid learning using data.

[0018] The AI ​​part above understands and converts the associations and correlations from the state of the objective facts, data, through the extraction of meaning and pattern recognition. After that, the resulting information internalized as unique knowledge is structured and turned into knowledge, and finally, by structuring knowledge, wisdom, which is a creative outcome, is produced.

[0019] The data collection section of the disaster safety knowledge base collects overseas information, unstructured data, and structured data, and stores them in the big data section of the disaster safety knowledge base through knowledge resource collection / management, human mimicry knowledge learning, and natural language understanding knowledge learning. In addition to natural language understanding knowledge learning, the big data section may also be provided with knowledge curation and compound reasoning knowledge augmentation.

[0020] The AI ​​section utilizes data accumulated in the big data section to make judgments and inferences using human cognitive abilities (language, speech, vision, sensitivity, etc.) and learning and inference functions. It is made up of a process of interpreting queries and analyzing problems by interpreting situations from queries and keywords provided from the outside, understanding the user's intentions, a process of searching for and inferring answer candidates, a process of making judgments and predictions through self-learning and growth, and selecting and generating answers, and finally a process of automatically generating deep query responses or analysis reports.

[0021] The disaster safety knowledge base is composed of a complex reasoning knowledge augmentation section, a natural language understanding knowledge learning section, a human mimicry knowledge learning section, and a knowledge resource collection / management section.

[0022] The complex reasoning knowledge augmentation unit is configured to extract rules and knowledge relations from structured and unstructured documents, and to search / infer new facts based on the extracted rules and knowledge relations to generate knowledge.

[0023] The natural language understanding knowledge learning unit is configured to analyze the structure, situation, context, and intent of a query input by a user and improve natural language understanding capabilities.

[0024] The human mimicry knowledge learning unit is configured to mimic human cognition and judgment functions and perform self-learning based on data accumulated in the knowledge base to improve performance. The knowledge resource collection / management unit is configured to collect and manage non-standard / standard data and various data from overseas.

[0025] The AI ​​unit mentioned above is a disaster safety knowledge bot capable of responding to in-depth queries via AI, and is composed of a user interface, a query / keyword input unit, a problem analysis unit, a user intent understanding unit, an answer candidate search / inference unit, an answer selection / generation unit, and a response generation unit.

[0026] The query / keyword input unit is composed of a query / keyword collection unit for collecting queries / keywords from a user, and a query identification unit for identifying contextual errors or typos in the query / keyword itself and requesting re-input or transmitting the query / keyword to the problem analysis unit.

[0027] The problem analysis unit is composed of a query interpretation unit for interpreting the sentence structure and words of the input query sentence, and a situation interpretation unit for interpreting the situation and context contained in the query.

[0028] The user intent understanding unit is composed of an intent extraction unit for extracting the user intent contained in the query sentence and transmitting it to the intent interpretation unit, and an intent interpretation unit for interpreting the user intent contained in the data received from the intent extraction unit.

[0029] The answer candidate search / inference unit is composed of an answer candidate search unit for searching answer candidates based on the analyzed query, and an answer candidate inference unit for inferring and ranking the optimal answer from the answer candidate list based on the user's intention and context.

[0030] The answer selection / generation unit is composed of an answer selection unit for selecting an optimal answer based on the ranked answer candidates, and an answer generation unit for generating an answer based on the selected answer.

[0031] The answer generation unit is composed of an answer implementation unit for generating answers in colloquial sentences that are easy for the user to understand, and an answer display unit for displaying the answers on a user interface. Effect of the Invention

[0032] In the AI-based disaster safety knowledge integrated management system of the present invention, by utilizing the sharing of disaster safety data and intelligent analysis services, it is possible to provide an AI-based question-answering service for specialized knowledge in the field of disaster safety, and an automatic reporting service that supports policy planning and report creation for specific subjects. In addition, the time, manpower, cost, and workload required for information search, analysis, and reporting can be dramatically reduced, so even non-experts and people with little experience can analyze and generate policy materials with the support of machines. Furthermore, comprehensive judgments can be made by taking into account data other than disasters, new technologies, and new issues, and the level of support for disaster safety policies can be significantly improved through an integrated knowledge base in the field of disaster safety. [Brief description of the drawings]

[0033] [Figure 1] 1 is a block diagram showing a schematic configuration of the present invention; [Diagram 2] 1 is a diagram showing an overview of a disaster safety knowledge consulting technology based on the present invention. [Diagram 3] FIG. 1 is a configuration diagram of an AI-based disaster safety knowledge integrated management system based on the present invention. [Figure 4] 1 is a diagram showing a process of disaster safety management based on user experience and intuition based on the conventional technology. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0034] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings, although it is to be understood that the scope of the present invention is not limited thereto.

[0035] The present embodiments are provided herein to fully disclose the present invention, and the scope of the present invention is defined only by the claims. Therefore, in some embodiments, well-known components, well-known operations, and well-known techniques are not specifically described to avoid obscuring the present invention.

[0036] The terms used in this specification are for the purpose of describing the embodiments and are not intended to limit the present invention in any way. In this specification, the singular form includes the plural form unless otherwise specified in the context. In addition, a component or operation referred to as "comprising" does not exclude the presence or addition of one or more other components or operations.

[0037] The present invention relates to an AI-based disaster safety knowledge integrated management system that utilizes disaster safety data sharing and intelligent analysis services of disaster safety data to enable AI-based question-answering services of specialized knowledge in the field of disaster safety and automatic reporting services that support policy planning and report material creation for specific subjects.

[0038] This invention can be broadly classified into two parts: a disaster safety knowledge base part connected to a data network, and an artificial intelligence part that uses artificial intelligence to realize high-level information processing by humans (see Figure 1).

[0039] The disaster safety knowledge base unit is composed of a data collection unit for collecting and aggregating various information from external organizations, a data transmission unit for sending the aggregated information to a server via LTE, 5G, WIFI, etc., and a big data unit for analyzing and storing the sent data.

[0040] The data collection unit mentioned above will provide general information such as disaster situations, disaster statistics, disaster policies, situation reports, and research reports; unstructured data such as news, social media, laws and regulations, situation reports, and research reports; and structured data such as damage recovery, weather, safety indices, disaster warnings, and 119 calls; but in some cases, data collection will be possible through IoT, with information exchange between all objects, such as machine-to-machine and machine-to-human, and the big data unit's ability to accumulate and analyze data will likely be strengthened as its information processing capabilities improve.

[0041] The artificial intelligence department mentioned above utilizes the data accumulated and analyzed by the big data department mentioned above to perform judgments and inferences based on human cognitive abilities (language, voice, vision, sensitivity, etc.) and learning and inference functions, making the machine intelligent through rapid learning using data and enabling it to create new value.

[0042] Meanwhile, the data collection unit may provide data through system linkages or cooperation between agencies. For example, stability inspection information may be provided through linkage with the National Safety Information Integrated Disclosure System, disaster management information through linkage with the National Disaster Management Information System, stability statistics through linkage with the Stability Information Integrated Management System, and safety report information through linkage with the Safety News Publicity System.

[0043] In addition, disaster case studies may be provided in cooperation with disaster response organizations, disaster case white papers may be provided in cooperation with white paper preparation organizations, and research reports may be provided in cooperation with disaster safety research organizations.

[0044] The big data section separates, analyzes, and accumulates the data provided by the data transmission section. It separates and analyzes real-time data such as steep slope sensors, water level information, and disaster accident site photos, structured data such as past damage history information and safety information for various facilities, and unstructured data such as disaster situation report text, disaster situation report images, and briefing materials.

[0045] The separated and analyzed data will then be classified into categories such as damage status of disasters and accidents, response history of disasters and accidents, and disaster safety policy information, and stored as archive data, making it possible to make the data public and shareable with those who need it (the general public, related organizations, etc.).

[0046] On the other hand, in order to utilize fragmented data as useful information, it is necessary to comprehensively explore data and experience, convert it into knowledge, use it in policy planning and disaster management, and gain insight into potential risk factors. By deriving meaning from the raw data, which is an objective fact, understanding the associations / correlations through pattern recognition, and converting it into information, and structuring and converting the resulting information contained as inherent knowledge, wisdom can ultimately be derived.

[0047] FIG. 2 is a diagram showing an overview of the disaster safety knowledge consulting technology according to the present invention. In the data collection unit of the disaster safety knowledge base, overseas information such as disaster situations, disaster statistics, and disaster policies, unstructured data such as news, SNS, and laws and regulations, and structured data such as damage recovery, weather, and safety indexes are collected, and are accumulated in the big data unit of the disaster safety knowledge base through knowledge resource collection / management or human mimicry knowledge learning and natural language understanding knowledge learning. In addition to natural language understanding knowledge learning, the big data unit also provides knowledge curation and complex inference knowledge expansion. In the artificial intelligence unit, human cognitive abilities (language, voice, vision, sensibility, etc.) and learning, judgment and inference are performed using inference functions by utilizing the data accumulated in the big data unit.

[0048] It consists of a process of understanding the user's intent through an externally provided query, query interpretation from keywords, and problem analysis through situation interpretation, a process of searching / inferring answer candidates, a process of selecting / generating answers through self-learning and growth, and finally a process of automatically generating a deep query response or analysis report.

[0049] FIG. 3 is a block diagram of the AI-based disaster safety knowledge integrated management system according to the present invention. The AI-based disaster safety knowledge integrated management system according to the present invention is a disaster safety knowledge bot that uses AI technology for language, learning, and inference to provide factual information such as disaster safety cases, experiences, and statistics, as well as to answer deep questions through question inference.

[0050] The disaster safety knowledge base is disaster safety knowledge data that utilizes big data and AI technology to support efficient information sharing in disaster situations with the aim of supporting database decision-making, and is composed of a base complex inference knowledge expansion part, a natural language understanding knowledge learning part, a human mimicry knowledge learning part, and a knowledge resource collection / management part.

[0051] The above-mentioned complex reasoning knowledge expansion unit extracts rules and knowledge relations from structured and unstructured documents, and based on this extracts / infers new facts to generate knowledge.

[0052] The natural language understanding knowledge learning unit mentioned above accumulates the results of analyzing the structure, situation, context, and intent of queries entered by users, and improves natural language understanding capabilities.

[0053] The Human Mimetic Knowledge Learning Unit mentioned above mimics human recognition and judgment functions and improves its performance through self-learning based on the data accumulated in the knowledge base.

[0054] The knowledge resource collection / management unit is configured to collect and manage unstructured / structured data and various overseas data.

[0055] On the other hand, the artificial intelligence unit can be considered a disaster safety knowledge bot capable of deep question answering through AI, and is composed of a user interface, a question / keyword input unit, a problem analysis unit, a user intent understanding unit, an answer candidate search / inference unit, an answer selection / generation unit, and an answer generation unit.

[0056] The question / keyword input unit is composed of a question / keyword collection unit for collecting questions / keywords from users, and a question identification unit for identifying contextual errors or typos in the question / keyword itself and requesting re-input or transmitting the question / keyword to the problem analysis unit.

[0057] The problem analysis section is composed of a question interpretation section for interpreting the sentence structure and words of the input question, and a situation interpretation section for interpreting the situation and context contained in the question.

[0058] The above-mentioned user intent understanding unit is composed of an intent extraction unit for extracting the user intent contained in the question sentence and transmitting it to the intent interpretation unit, and an intent interpretation unit for interpreting the user intent contained in the data received from the intent extraction unit.

[0059] The answer candidate search / inference unit is composed of an answer candidate search unit for searching answer candidates based on the analyzed question, and an answer candidate inference unit for inferring and ranking the most suitable answer from the answer candidate list based on the user's intention and context.

[0060] Homophones require different answers depending on the user's intent and context.

[0061] The answer selection / generation unit is composed of an answer selection unit for selecting an optimal answer based on the ranked answer candidates, and an answer generation unit for generating an answer based on the selected answer.

[0062] The answer generation unit is composed of an answer implementation unit for generating answers in colloquial sentences that are easy for the user to understand, and an answer presentation unit for displaying the answers on a user interface.

[0063] In addition, the AI ​​department can be further configured with a report generation department, a knowledge-based language generation department, and a content processing and generation department to automatically assist in the creation of policy planning and report materials, and provide services to support decision-making and reduce time and manpower consumption.

[0064] The report generation unit is composed of a template generation and content synthesis unit for generating report content in accordance with a template format, and a template and output interface unit for providing the report content generated thereby in accordance with the template format so that the user can easily understand it.

[0065] The knowledge base language generation unit is composed of a data extraction and support search unit for extracting and searching necessary data from the disaster safety knowledge base, and a content configuration and data catalogue for constructing content using the data extracted and searched by this unit.

[0066] The content processing and generation unit is composed of a table / graph and text processing unit for generating statistical data (tables / graphs) based on text, and an automatic title and content generation unit for generating content in a format that is easy for users to understand and extracting and summarizing the main content.

[0067] Therefore, the artificial intelligence unit can utilize the data accumulated and analyzed in the big data unit to gradually execute judgments and inferences based on human perceptual abilities such as language, voice, vision, and sensitivity, as well as learning and inference functions.

[0068] In addition, the question / keyword input section of the artificial intelligence section performs deep question answering by question answering / keyword search using the chatbot driver. Along with the chatbot driver, input device, management system, and machine learning tool, it is composed of a question answer pre-management module, a user intention understanding module, a dialogue agent, and a dialogue management system module. The question answer pre-management module manages the contents sent by the Intent Finder and Dialog Agent through session control.

[0069] A session is generated using four conditions: USER, DEVICE, CHATBOT, and a certain period of time.

[0070] The above-mentioned user intent understanding module is configured to understand the user intent, send a message to the most suitable DIALOG AGENT, understand the dialogue intent based on various features, and then support the search for an answer.

[0071] The above-mentioned dialogue agent system is configured to respond to user utterances in a dialogue agent platform environment, to be developed and distributed in a disaster safety data environment, and to be registered and executed through ADMI.

[0072] The dialogue management system described above is configured to enable various dialogues, such as daily conversations and news, depending on the type of Q&A engine and knowledge base, and to enable the definition of slots and tasks in the SDS scenarios generated by the dialogue modeling tool.

[0073] The present invention includes a function for semi-automatic or automatic generation of reports. The report design and report server part is configured to generate reports via data links and provide report integration / distribution and report utilization procedures.

[0074] Each report created based on various report creation sources (such as knowledge bases) linked with the report agent is integrated and distributed on the distribution server side, and the report contents reflected in the content management system of the application server are finally transferred to the report server and made available to users.

[0075] As described above, the AI-based disaster safety knowledge integrated management system of the present invention utilizes disaster safety data sharing and disaster safety data knowledge analysis services to enable an AI-based question-answering service for specialized knowledge in the field of disaster safety, and an automatic report service that supports policy planning and the creation of report materials for specific topics.

[0076] In addition, it can dramatically reduce the time, manpower, costs, and workload required for data exploration, analysis, and reporting, and even non-experts or people with relatively little experience can perform analysis and generate policy materials with the assistance of machines. It also helps to enable comprehensive judgments by taking into account data other than disasters, new technologies, and new issues, thereby significantly improving the level of support for disaster safety policies through insights from an integrated knowledge base in the field of disaster safety.

[0077] Although the technical idea of ​​the present invention has been specifically described in the preferred embodiment, please note that the above embodiment is for the purpose of explanation and not for the purpose of limitation. It is clear that various modifications and alterations are possible within the scope of the technical idea of ​​the present invention, and therefore, it is natural that these modifications and alterations are included in the scope of the claims.

Claims

1. This is an AI-based disaster safety knowledge integrated management system that will enable AI-based question-answering services on specialized knowledge in the field of disaster safety by sharing disaster safety data and utilizing intelligent analysis services, and will enable automated reporting services to support policy planning and report creation on specific topics. It consists of a disaster safety knowledge base part connected to a data network and an AI part for realizing high-dimensional information processing using AI. The disaster safety knowledge base unit is composed of a data collection unit for collecting and aggregating various information from external organizations, a data transmission unit for transmitting the collected information to a server via LTE, 5G, WIFI, etc., and a big data unit for analyzing and storing the transmitted data. The big data unit classifies, analyzes and stores the data provided by the data transmission unit, and separates and analyzes real-time data such as steep slope sensors, quantity water level information, disaster accident site photos, etc., regular data such as past damage history information, safety information of various facilities, etc., and non-regular data such as disaster situation report text, disaster situation report image, briefing materials, etc. The data thus classified and analyzed is classified into disaster accident damage situation, disaster accident response history, and disaster safety policy information, and stored as archive data. The AI ​​department utilizes the data accumulated and analyzed by the Big Data department, and performs judgment and inference based on human cognitive abilities (language, voice, vision, sensitivity, etc.) and learning, and inference functions in stages to make the machine intelligent. The AI ​​unit is a disaster safety knowledge bot capable of deep question answering using AI, and is composed of a user interface, a question / keyword input unit, a problem analysis unit, a user intent understanding unit, an answer candidate search / inference unit, an answer selection / generation unit, and an answer generation unit. The question / keyword input unit of the artificial intelligence unit performs deep question answering by question answering / keyword search using a chatbot driving device, and is composed of a question answer pre-management module, a user intention understanding module, a dialogue agent and a dialogue management system module, together with a chatbot driving device, an input device, a management system and a machine learning tool. The question and answer pre-management module manages the contents of Intent Finder and Dialog Agent through session control, and the session is generated using four conditions: USER, DEVICE, CHATBOT, and a certain range of time. The user intent understanding module understands the user intent and sends a message to the most appropriate Dialog Agent. It is configured to be able to support answer search after understanding the dialogue intent based on various features. The dialogue agent system responds to user utterances in a dialogue agent platform environment, is developed and distributed in a disaster safety data environment, and is configured to be registered and executed via ADMI. The dialogue management system is capable of various dialogues such as daily dialogues and news depending on the type of Q&A engine and knowledge base, and is configured to allow slot and task definition in the SDS scenario generated by the dialogue modeling tool. This is an AI-based disaster safety knowledge integrated management system.

2. In paragraph 1, the artificial intelligence department extracts meaning from raw data, which is objective fact, understands associations / correlations through pattern recognition, turns it into information, and encapsulates it as unique knowledge. This information is then structured and turned into knowledge, and finally, the system derives the structuring of knowledge. This is an integrated disaster safety knowledge management system using AI.

3. In paragraph 1, the data collection section of the disaster safety knowledge base collects overseas information, unstructured data and structured data, and stores the collected data in the big data section of the disaster safety knowledge base through knowledge resource collection / management or human mimicry knowledge learning and natural language understanding knowledge learning. This is an integrated disaster safety knowledge management system using AI, characterized in that the big data section provides not only natural language understanding knowledge learning, but also knowledge curation and complex inference knowledge expansion.

4. In paragraph 1 or 2, the artificial intelligence unit utilizes the data accumulated and analyzed by the big data unit to perform judgment and inference using human cognitive abilities (language, voice, vision, sensitivity, etc.) and learning and inference functions, The process of understanding the user's intent by analyzing the query provided from an external source, interpreting the query from keywords, and analyzing the problem by interpreting the situation; the process of searching / inferring answer candidates; An AI-based disaster safety knowledge integrated management system that is characterized by being composed of processes of self-learning and growth, judgment / prediction, and answer selection / generation; and finally, a process of deep query response or automatic generation of analysis reports.

5. In paragraph 1 or 3, the disaster safety knowledge base is an AI-based disaster safety knowledge integrated management system consisting of a complex inference knowledge enhancement unit, a natural language understanding knowledge learning unit, a human mimicry knowledge learning unit, and a recognition resource collection and management unit, as disaster safety knowledge data that aims to support database decision-making and supports efficient information sharing in disaster situations by utilizing big data and AI technology.

6. In the fifth paragraph, the composite reasoning knowledge augmentation unit is configured to extract rules and knowledge relationships from structured and non-structured documents, and to search for / infer new facts based on the extracted rules and knowledge relationships to generate knowledge.

7. In the fifth paragraph, the natural language understanding knowledge learning unit is configured to accumulate the results of analyzing the structure, situation, context, and intent of a question sentence input by a user, and to improve natural language understanding capability.

8. An AI-based disaster safety knowledge integrated management system, characterized in that the question / keyword input unit in paragraph 1 is composed of a question / keyword collection unit for collecting questions / keywords from users, and a question recognition unit for identifying contextual errors or typos in the question / keyword itself and requesting re-input, or transmitting the question / keyword to the problem analysis unit.

9. An AI-based disaster safety knowledge integrated management system, characterized in that the problem analysis unit in paragraph 1 is composed of a question interpretation unit for interpreting the sentence structure and words of the input question, and a situation interpretation unit for interpreting the situation and context contained in the question.

10. The AI-based disaster safety knowledge integrated management system in paragraph 1, wherein the user intent understanding unit is comprised of an intent extraction unit for extracting the user intent contained in the question sentence and transmitting it to the intent interpretation unit, and an intent interpretation unit for interpreting the user intent contained in the data received from the intent extraction unit.

11. In paragraph 1, the answer candidate search / inference unit is characterized in that it is composed of an answer candidate search unit for searching answer candidates based on the analyzed question, and an answer candidate inference unit for inferring and ranking the optimal answer from the answer candidate list based on the user's intention and context.

12. In paragraph 1, the answer selection / generation unit is characterized in that it is composed of an answer selection unit for selecting the optimal answer based on ranked answer candidates, and an answer generation unit for generating an answer based on the selected answer.

13. In paragraph 1, the answer generation unit is characterized in that it is composed of an answer implementation unit for generating answers in colloquial sentences that are easy for users to understand, and an answer display unit for transmitting the answers to a user interface.

14. In paragraph 1, the artificial intelligence unit of the AI-based disaster safety knowledge integrated management system is characterized in that it further comprises a report generation unit, a knowledge-based language generation unit, and a content processing and generation unit in order to provide services that are useful for decision-making support and reduce time and manpower consumption through automatic support for policy planning and report materials.

15. In paragraph 14, an AI-based disaster safety knowledge integrated management system is characterized in that the report generation unit is composed of a template generation and content synthesis unit for generating report content according to a template format, and a template and output interface unit for providing the report content generated thereby according to the template format so that the user can easily understand it.

16. In paragraph 14, an AI-based disaster safety knowledge integrated management system is characterized in that the knowledge base language generation unit is composed of a data extraction and support search unit for extracting and searching necessary data from the disaster safety knowledge base, and a content configuration and data catalog for constructing content using the extracted and searched data.

17. In paragraph 14, an AI-based disaster safety knowledge integrated management system, characterized in that the content processing and generation unit is composed of a table / graph and text processing unit for generating statistical data (tables / graphs) on a text basis, and an automatic title and content generation unit for generating content in a format that is easy for users to understand and extracting and summarizing the main content.

18. In Section 4, the report design and report server part is configured to generate reports through data links and provide report integration / distribution and report utilization procedures, and each individual report created based on various report creation sources (knowledge base, etc.) linked with the report agent is integrated and distributed on the distribution server side, and the contents of the report reflected in the content management system of the application server are finally transferred to the reporting server and made available to users, in this AI-based disaster safety knowledge integrated management system.

Citation Information

Patent Citations

  • Disaster prevention system

    JP2021114150A

  • A freshness maintaining device and a method for maintaining freshness using the same

    KR102646425B1

  • Machine learning-based relationship association and related discovery and search engines

    US20190354544A1