A student in Carnegie Mellon University’s School of Computer Science (SCS) has used reinforcement learning to help control nuclear fusion reactions, a significant step toward harnessing the immense power produced in nuclear fusion as a source of clean, abundant energy.