Artificial intelligence (AI) has made significant progress in recent years, but physicists have been slow to utilize it in their search for fundamental laws of nature. This is not because they are afraid of losing their jobs, but rather because AI tends to provide “what” rather than “why” answers, which are not particularly useful for theory development. However, a new approach has been developed that allows deep learning algorithms to express themselves in the language of physics, enabling them to analyze large data sets and identify meaningful patterns and equations. Steve Brunton at the University of Washington in Seattle notes that this marks a move into the “discovery phase” for AI in physics research.
