Machine learning (ML) and artificial intelligence (AI) are transforming many fields of science, and interest in applying these methods to the problems of astrobiology is accelerating. This includes…

Arizona State University received a $939.7K cooperative agreement from NASA's Science Mission Directorate (awarded December 1, 2024, with completion by November 30, 2027) to develop critical data infrastructure supporting machine learning and artificial intelligence applications in astrobiology research. Under solicitation 24-SMDSS24-0090, ASU will create two benchmarked datasets—cheminformatic and spectrometric data—empirically annotated with ground truth labeling to enable training of ML/AI algorithms for life detection missions. The project addresses a fundamental infrastructural gap in NASA's astrobiology efforts by establishing standardized protocols and data handling standards that will support the broader community's development of standardized data resources beyond this initial project, positioning ML/AI approaches to advance the agency's mission science objectives in the search for life on other planets.

The award carries no set-aside designation, reflecting ASU's competitive positioning as a leading research institution. ASU will perform the work at its Tempe, Arizona location through its Office of Research and Sponsored Projects Administration. This cooperative agreement aligns with NASA's broader investment in AI and astrobiology integration, as evidenced by the agency's public resource initiatives for enabling community training and tool sharing. The project builds upon ASU's substantial track record executing complex research and development initiatives for federal agencies, including prior NASA Science Mission Directorate awards focused on science activation and balanced science programs spanning astrophysics, planetary science, and related disciplines.

Machine learning (ML) and artificial intelligence (AI) are transforming many fields of science, and interest in applying these methods to the problems of astrobiology is accelerating. This includes efforts to support the application and development of ML/AI tools across astrobiology, with efforts underway to enable training and sharing of resources within the community using ML/AI tools (see e.g., https://WWW.NASA.GOV/about-a-i-astrobiology/). Success of these burgeoning ML/AI efforts for life detection will require infrastructure also to be built supporting a public resource of high quality, standardized data resources for training ML/AI algorithms. This is a critical infrastructural need for NASA's life detection efforts that we aim to fill in this project. Our goal is to build two benchmarked datasets, specifically designed for the ML/AI needs of astrobiologists interested in life detection: these include, (1) cheminformatic and (2) spectrometric data, empirically annotated with ground truth labeling, and standards and protocols for generating and handling data that will allow further growth of the community's standardized data resources beyond this initial project. These data infrastructure modules will fill the need for standardization and benchmarked data, necessary for ML/AI approaches to advance mission science in the search for life on other planets.