An AI strategy proven adept at board games like Chess and Go, reinforcement learning, has now been adapted for a powerful protein design program. The results show that reinforcement learning can do ...