Foundation models - artificial intelligence systems trained on massive data sets to perform a wide range of tasks - have the potential to transform scientific discovery and innovation. At the request of the U.S. Department of Energy (DOE), the National Academies conducted a study to consider the capabilities of current foundation models as well as future possibilities and challenges. Foundation Models for Scientific Discovery and Innovation explores how foundation models can complement traditional computational methods to advance scientific discovery, highlights successful use cases, and recommends strategic approaches and investments to support DOE's mission.
