MIT researchers have proposed an AI training method that uses optimized synthetic climate data to accelerate modeling and predict unknown emission scenarios. The new algorithm requires only one expensive Earth system model simulation to generate training data, and the resulting simulator can handle scenarios such as regional pollution controls that fall outside the training set.
Source:Electronics For You · Read Original
Leave a Reply