Cambridge researchers published a study in the journal PNAS about the development of an artificially intelligent algorithm called IcePic. The algorithm is capable of outperforming scientists in predicting how and when different materials produce ice crystals. It may assist atmospheric scientists in improving climate change predictions in the future.
How Artificial Intelligence Helps in Understanding Ice Formation
Water expands when it freezes. Yet, understanding water and how it freezes around different molecules has far-reaching implications, from weather systems that can affect entire continents to keeping biological tissue samples in a hospital.
The Celsius temperature scale was designed with the assumption that it is the transition temperature between water and ice. Yet, while ice always melts at 0°C, water does not always freeze at that temperature, as water can remain liquid at -40°C. It is also found that the impurities in water allow ice to freeze at higher temperatures.
An Artificial Intelligence Algorithm Measures Ice Nucleation
One of the field's primary goals has been to anticipate various materials' ability to encourage ice production, often known as a material's ice nucleation ability.
The concept of ice nucleation begins with understanding what an ice nucleus is. According to Wikipedia, an ice nucleus, also known as an ice nucleating particle, which is present in the environment and serves as the nucleus for the development of an ice crystal. In the atmosphere, ice nuclei can catalyze the creation of ice particles by various methods of ice nucleation.
Water vapor can deposit directly on solid particles in the high troposphere. Ice nuclei can cause droplets to freeze in clouds warmer than 37 °C, where liquid water can remain in a supercooled condition. Contact nucleation can occur when an ice nucleus collides with a supercooled droplet, but the more important freezing mechanism occurs when an ice nucleus becomes submerged in a supercooled water droplet and then freezes. In the absence of an ice nucleating particle, pure water droplets can be supercooled to temperatures approaching 37 °C before freezing homogeneously.
Artificial Intelligence VS Human Researcher in Predicting Ice Nucleation
Researchers at the University of Cambridge created a deep learning program capable of predicting the ice nucleation capabilities of various materials. The tool has outperformed scientists through an online quiz in which they were asked to guess when ice crystals would form.
Deep learning is the process by which artificial intelligence (AI) learns to derive insights from raw data. It discovers its own patterns in the data, removing human intervention and allowing it to analyze results more quickly and precisely. In the instance of IcePic, it can infer various ice crystal formation parameters around various materials. IcePic has been trained on thousands of photos to examine completely new systems and make reliable predictions.
The researchers devised a quiz in which scientists were asked to predict when ice crystals would develop under various conditions depicted by 15 different photographs. The outcomes were then compared to IcePic's performance.
When tested, IcePic outperformed over 50 researchers from across the world in determining a material's ice nucleation ability. It also helped in pinpointing the areas where people were making mistakes. The study of ice formation has become increasingly crucial in climate change studies.
Water travels continuously inside the Earth's atmosphere, condensing to form clouds and falling in the form of rain and snow. Different foreign particles, such as pollution-related smoke particles, influence how ice forms in these clouds.
"The nucleation of ice is really important for the atmospheric science community and climate modeling. At the moment, there is no viable way to predict ice nucleation other than direct experiments or expensive simulations. IcePic should open up a lot more applications for discovery," said the first author of the study Michael Davies.
RELATED ARTICLES : Ice Formation on Graphene Shows the Need for Heat
Check out more news and information on Technology in Science Times.
