Machine Learning Model To Predict Disastrous Events

Researchers have developed a machine learning model that can predict future natural disasters such as earthquakes and pandemics.

The world is a chain of events. Hopefully the future life or death and even events. Many studies suggest that events – a major disaster – occur at a certain interval of time. “Statistically speaking, these events are so rare that there isn’t enough data on them to use predictive models to accurately predict when they will happen next.” Researchers at Brown University and MIT believe it can be predicted. What if this time can be calculated and we can prepare before the disaster occurs?

The researchers describe how they combined statistical algorithms—which require little data to make accurate, efficient predictions—with a powerful machine learning technique developed at Brown and trained them to predict the scenarios, probabilities and sometimes even the timeline of extraordinary events despite the lack of historical record. to them.

They found that this framework provides a way to avoid the need for large amounts of data required for computation. They use a sequential sampling technique called active learning. To study big data, they used DeepOnet, a type of artificial neural network, which uses interconnected nodes in successive layers that closely mimic the connections made by neurons in the human brain. .

Researchers have applied pinpointing parameters and different ranges of probabilities for dangerous spikes during a pandemic, finding and predicting rogue waves, and predicting when a ship will break due to stress. For example, in rogue waves—which are more than twice the size of surrounding waves—researchers have found that they can detect and quantify when rogue waves form by looking at potential wave conditions that do not interact linearly with time, leading to waves sometimes three times. their original size.

The researchers found that their new method outperformed more traditional modeling efforts, and they believe it presents a framework that can effectively discover and predict all kinds of unusual phenomena.

Reference: Ethan Pickering, Discovering and predicting extreme events through active learning neural operators, Nature Computational Science (2022). DOI: 10.1038/s43588-022-00376-0. www.nature.com/articles/s43588-022-00376-0