MIT's New Algorithm Generates Extreme Event Scenarios Without Extreme Data

It's a machine learning algorithm developed by MIT researchers that can generate scenarios for rare and extreme events without relying on abundant historical data on those events, by understanding the underlying patterns of complex systems.

Traditional methods rely on historical data of similar events, while the new algorithm learns from available data (even if it doesn't include the extreme event) to generate 'what-if' scenarios for unprecedented events.

The region is exposed to shocks like energy price volatility, drought, or global supply chain disruptions. This algorithm allows governments and companies to stress-test their infrastructure and supply chains against scenarios that haven't happened yet, enhancing proactive planning.

The research is advanced but still within the academic sphere. Practical application would require collaboration with infrastructure or risk management entities to adapt it to local data.