India, Nov. 7 -- If teaching machines to understand the physical world is the next frontier of artificial intelligence, Caltech professor Anima Anandkumar has done more than most to map the route. Can an algorithm forecast the weather faster than a supercomputer? Can AI help engineers design a safer medical device? Can a machine learn how nature behaves from data alone? Some of the most influential answers have come from her work on neural operators, a class of AI models she pioneered that learn how physical systems evolve and run orders of magnitude faster than conventional simulations. She has used them to help build high-resolution weather models and to design catheters that resist infection. Prior to these breakthroughs, she had made he...