Machine learning, which is able to model complex systems and environments far beyond what can be done with traditional software, is leading to new advances in the aerospace and defense industry. In space exploration, for instance, space debris is a growing and serious problem, with the risk of collision increasing as growing numbers of satellites are deployed. Currently, the amount of debris littering in orbit is unknown, with estimates for pieces bigger than 1cm ranging from 900,000 to 160 million. But machine learning may change that uncertainty, by the use of self-learning algorithms to improve avoidance maneuvers, forecast potential collisions in space, and even support space junk removal.