Computer scientists have developed an artificial intelligence (AI) program capable of predicting when catastrophic tipping points will occur and aim to use it for forecasting ecological collapse, financial crises, pandemics, and blackouts.
The researchers published their findings in the journal Physical Review X.
AI Predicts Accurately
Tipping points are sudden changes after which a local system or its environment shifts to an undesirable state that is difficult to recover from. For example, if the Greenland ice sheet collapses, it will reduce snowfall in the northern part of the island, significantly raise sea levels, and render many parts of the ice sheet irretrievable.
However, the science behind these powerful transitions remains poorly understood and often relies on overly simplistic models, making accurate predictions challenging. Previously, scientists used statistics to assess the declining strength and resilience of systems through their increasing fluctuations.
A photo of a forest fire. (Photo: Jackal Yu|).
To find a more accurate way to predict dangerous transitions, researchers combined two different types of neural networks or algorithms that simulate how information is processed in the brain. The first type breaks down complex systems into large networks of interacting nodes before tracking the connections between them; and the second monitors how individual nodes change over time.
Since tipping points are difficult to predict, knowing where to find them is equally challenging, making real data on sudden critical transitions scarce. To train their model, the researchers turned instead to tipping points in simple theoretical systems—such as modeled ecosystems and asynchronous counting machines, which, given enough time, begin to move together. The researchers reported that AI accurately predicted what would happen.
Decoding the Black Box of Wildfires, Pandemics, and Financial Crises
A challenge in predicting human-related systems is understanding and responding to their own forecasts.
For example, in urban traffic: while it may be easy to identify congested routes, relaying real-time congestion information can lead to chaos. Drivers may immediately change their routes, which can reduce congestion on some roads but simultaneously create congestion on others. This dynamic interaction makes predictions particularly complex.
To address this issue, the researchers indicated they would focus on parts of the human system that appear unaffected. In the case of road networks, this could be done by examining routes that are congested due to their basic design.
Utilizing AI to capture these underlying signals is valuable for making predictions. Although predicting such systems is challenging, it is worthwhile because critical transitions in human-involved systems can lead to even more severe consequences.
