7+ hour, 7+ min ago  (670+ words) MIT researchers have developed a novel machine-learning framework that significantly improves the accuracy and efficiency of simulating chemically disordered metal alloys. Published recently in Science Advances, the work addresses a longstanding bottleneck in materials science: predicting how complex atomic arrangements…...