A new algorithm has been developed to learn and predict unprecedented scenarios that critical infrastructure and global supply chains are least prepared for.
This approach enables the anticipation of extreme events without requiring prior examples of such events, addressing a key challenge in risk modeling.
The ability to generate these scenarios can improve planning and resilience for systems vulnerable to rare but impactful disruptions.
Source: https://news.mit.edu/2026/generating-scenarios-extreme-events-without-extreme-data-0824