Currency
Artificial intelligence will be able to predict 'tipping points' of future disasters: what is known
Predicting dangerous tipping points in complex systems has previously been a daunting task for scientists. But now, thanks to a new artificial intelligence system, that job can be done for them.
Computer scientists have created an artificial intelligence (AI) program that can anticipate the onset of catastrophic tipping points. Scientists plan to use it to predict environmental collapse, financial collapses, pandemics and blackouts, Live Science writes.
Turning points are sudden shifts after which a localized system or its environment goes into an undesirable state from which it is difficult to return. For example, if the Greenland ice sheet were to fall, it would also reduce snowfall in the northern part of the island, dramatically raise sea levels, and make large portions of the shield irretrievable.
However, the science behind these possible events is poorly understood and often based on simplistic models, making accurate predictions difficult. Previously, scientists have used statistics to estimate the decreasing strength and stability of systems from their increasing vibrations. But the results of studies using such statistical methods are contradictory.
So to find a more accurate way to anticipate dangerous transitions, the researchers combined two different types of neural networks, or algorithms that mimic the way the brain processes information. The first broke down complex systems into large networks of interacting nodes before tracking the connections between nodes, while the second followed how individual nodes changed over time.
Because tipping points are difficult to predict, it's equally difficult to know where to look for them, making actual data on sudden critical transitions scarce. To train their model, the researchers turned to tipping points in simple theoretical systems, including model ecosystems and unsynchronized metronomes that start oscillating together in sufficient time.
"If a future critical transition can be predicted, then we can prepare for the shift, or perhaps even prevent the transition, and thus mitigate the effects. This prompted us to develop an artificial intelligence approach to anticipate the onset of such sudden transitions long before it happens," says senior study author Gang Yan, a professor of computer science at Tongji University in China.
Scientists specify that although it is difficult to foresee such systems, it is worth considering, as the use of artificial intelligence can be valuable for prediction.
Only verified information from us in Telegram-channel OBOZ.UA and Viber. Do not fall for fakes!