Shingo Yamaguchi Laboratory, Faculty of Environment and Information Studies, Keio University, National Institute of Information and Communication Sciences, and National Institute of Disaster Prevention Science and Technology are aiming to utilize artificial intelligence (AI) in the field of disaster prevention, and social networking in the event of a disaster using AI. We have compiled training guidelines for service (SNS) information analysis.The aim is to enable local governments and others to effectively carry out information analysis training in the event of a disaster using AI, and it is said that it will be possible to streamline and speed up situational judgments in the event of a disaster.
According to Keio University, this guideline is a provisional version.Chapter 1 provides an overview of the SNS information analysis system, and Chapter 2 provides a method for implementing training on a responsive diagram. Digitally processing SNS posts, which are expected to range from hundreds of thousands to millions. It makes it possible to grasp important information such as lifeline troubles, shortage of supplies at evacuation centers, and isolated victims.
By experiencing practical training on a regular basis, disaster information can be processed accurately in a short time and respond quickly when a disaster actually occurs.As a result, residents can be evacuated smoothly, and it is expected that the burden on local government officials will be reduced.
In the Great East Japan Earthquake, the victims posted 3,300 disaster information on Twitter alone on the day of the outbreak, but it was difficult for a limited number of local government officials to select important information from the vast amount of information.For this reason, the universities established a joint research conference (disaster prevention AI joint research conference) in June 2017, and have begun to formulate training guidelines.
