A new framework allows AI models that have already been trained to learn new tasks without sacrificing performance when performing old tasks. The framework, called CHEEM, also improves an AI model’s operating efficiency by using fewer computational steps to perform simpler tasks.

For decades, scientists have tracked a strange pattern deep beneath the Pacific Ocean: a fault line that produces nearly identical earthquakes every few years, almost like clockwork — something that’s rarely seen in nature. New research published in Science and led by Indiana University seismologist Jianhua Gong finally explains why.