Developing an AI model that accurately predicts changes in Arctic sea ice a year later.

Korean researchers have developed an artificial intelligence model that can predict changes in Arctic sea ice up to a year later with high accuracy.

A team led by Professor Lim Jeong-ho of the Department of Global Environmental City Construction Engineering at UNIST (UNIST) has developed artificial intelligence that can predict Arctic sea ice concentration with an error accuracy of less than 6% within a year.

The developed AI used the deep learning algorithm UNET to learn complex relationships between changes in past Arctic sea ice concentrations and major climate factors such as temperature, water temperature, and solar radiation.

The researchers explained that while the existing model recorded an average prediction error of 17.35%, the model recorded an average prediction error of 7.07%, reducing the error value by less than half, and that it is expected to help explore marine resources and establish policies to respond to climate change.