To reduce the risk of floods and forest fire-how to use mechanical learning for Google and AWS

　Google and Amazon Web Services (AWS) have revealed their efforts, such as machine learning models, which could support countries that are increasing around the world.

　The United Nations Change of the UN Subsection Treaty was held from the end of October to November 12, local time.Meanwhile, the two companies have introduced initiatives to reduce the risk of disasters due to climate change such as floods and forest fires.

flood

　Google has published a paper -free paper on the flood forecast system using a machine learning model.The system can provide "accurate real -time flood warnings" to related organizations and citizens, focusing on large -scale rivers with observation data.The paper was written by a researcher at Google Research and Jerusalem Hebrew University in Israel.

　Google's flood forecasts in 2018 send alerts to smartphones of people living in areas affected by floods.This is a part of Google's Crisis Supposer program, working with people working on site and rescue teams to develop technical development.

　The target area of the program has been expanded to most India and Bangladish in three years since 2018, covering areas where 220 million people live.By the time of the 2021 monsoon season, the target areas were further expanded and the area where 360 million people lived were covered.

　Google's engineering and crisis response vice president Yossi Matias said in a blog post, "With the improvement of flood forecasting technologies, more than 115 million alerts were sent. It was previously sent.It is three times the amount of alert. "

　Google alerts are notified how many centimeters of river water levels have risen, but that's not the only one.The "Google Map" layer can provide a "flooded map" that displays the range and depth of the flood.With the evolution of the company's new flooding models and all models in general, it has been significantly scaled up and provides more information to more people.

　Google researchers said, "Although there have been studies that have shown promising results so far, we can use machine learning as a core component and calculate accurate flood warnings at the right time.There were few possible systems. "