Catastrophes are difficult to predict because they are so rare. But AI using active learning can make predictions from very small data sets.

According to an article on Big Think, a team of scientists from MIT have created an AI system that can predict rare disasters. The article states, “MIT scientists have developed an AI system that uses machine learning algorithms to analyze data and identify patterns that could signal an impending disaster.”

Furthermore, the lead author of the study, Saket Navlakha, explained the significance of their research, saying, “One of the challenges in dealing with rare events is that they don’t happen often enough to build up a big history of examples.” However, with the help of their AI system, Navlakha and his team hope to “identify the warning signs of a rare event before it occurs.”

This technology has the potential to be incredibly valuable in fields such as finance, where sudden market crashes can have disastrous consequences. As Navlakha noted, “The ability to anticipate rare but consequential events is becoming increasingly important as we become more globally interconnected.”

Via The Impactlab