Scientists used AI to crack one of water's biggest mysteries

Scientists at the University of Osaka have leveraged AI to tackle one of water's longstanding enigmas, especially when it's supercooled. By training an AI model on computer simulations, the researchers sifted through 16 different structural descriptors to pinpoint the best methods for distinguishing between water's two competing liquid states. This breakthrough offers a clearer framework for understanding water's unusual properties, which could have far-reaching implications for fields like materials science and climate modeling. The ability to better describe water's microscopic structure could lead to advances in predicting weather patterns and developing new technologies.