Google, AWS Take Disaster Warning To The Next Level With Machine Learning

Google and Amazon Web Services (AWS) have ramped up efforts to tackle climate change effects by controlling floods and wildfires through machine learning. The companies demonstrated their initiatives during the United Nations Climate Change Conference (COP26) held in Glasgow, United Kingdom from 31 October – 12 November 2021.

Google in collaboration with Hebrew University published a research paper about its operational flood forecasting system that uses machine learning to give accurate real-time flood warnings to agencies and the public. According to Google, the system which became operational in 2018, has been successfully tested in India and Bangladesh which are prone to floods.

Google’s Vice President of Engineering and Crisis states that the flood forecasting initiative which began in 2018, was expanded in 2021 to cover much of India and Bangladesh.

“In 2021, our operational systems were further expanded to cover an area with over 360 million people. Thanks to better flood prediction technology, we sent out over 115 million alerts,” he said.

Google alerts, through its new machine-learning models that use Long Short-Term Memory (LTSM) deep neural networks can now show both the extent and depth of flooding as a layer on Google Maps.

“While previous studies provided encouraging results, it is rare to find actual operational systems with ML models as their core components that are capable of computing timely and accurate flood warnings,” Google’s researchers said.

On the other hand, AusNet, an energy company based in Australia has been using Google’s LiDAR cameras and AWS SageMaker machine learning to map out the state’s vegetation areas that need to be trimmed to stem bushfire threats.

According to AusNet its collaboration with AWS has increased worker safety by using LiDAR data and reducing the need for engineers, surveyors, and designers to travel to sites.

The energy firm adds that the collaboration has “resulted in 80.53% accuracy across all five segmentation categories, saving AusNet an estimated $366,627 per year through automated classification”.

AusNet manages 54,000 kilometres of power lines and brings energy to more than 1.5 million Victorian homes and businesses. 62% of this network is located in high bushfire risk areas.