Machine learning can help us predict landslides caused by climate change

Although landslides are natural phenomena, their occurrences are influenced by human activities. In the US, landslides are becoming more frequent and severe. Christoph Mertz of Carnegie Mellon’s Robotics Institute is attempting to use deep learning algorithms to detect signs of impending landslides. His project requires a lot of data and cross-discipline collaboration. Experts are needed in computer science, geology, infrastructure, and other fields. Ultimately, Mertz hopes not just to predict and prevent landslides, but also to bring about infrastructure improvements that are related to the issue.

Key Takeaways:

Landslides are serious problems resulting in annual deaths of 25 to 50 people, and billions in damage.
Machine learning develops algorithms to analyze pictures of infrastructure to accurately predict if a landslide is probable in that area.
The leader of the initiative is Christopher Mertz who works at Carnegie Mellon University’s Robotics Institute.

“Landslides are natural phenomena, but many of the conditions that can increase their likelihood are caused by human activity, such as directing surface runoff to an area or altering natural slopes for the construction of buildings and roads”