Frontier Development Lab (FDL) kicks off its sixth year, with teams comprised of researchers who will use AI and ML to tackle seven challenges in the areas of heliophysics, astronaut health, lunar resources and Earth science. FDL applies AI technologies to science to push the frontiers of research and develop new tools to help solve some of humanity's biggest challenges. For the second year in a row, FDL, which typically brings researchers, mentors and faculty from around the world together at the SETI Institute in Mountain View, CA, will take place virtually.
“In an impressive pivot, our 2020 FDL participants demonstrated that interdisciplinary researchers could achieve extraordinary results in an intense sprint environment and do it virtually, across about nine time zones,” said Bill Diamond, President and CEO of the SETI Institute. “We The FDL AI/ML research accelerator will again be a virtual program this year, but we anticipate extraordinary results!”
FDL is a public-private partnership with NASA in the USA and ESA in Europe. It brings together some of the brightest minds from space science, AI and the commercial sector, including Google Cloud, Lockheed Martin, Luxembourg Space Agency, Intel, Microsoft, MIT Portugal, Mayo Clinic, USGS and NVIDIA. New partners this year include Lawrence Berkeley National Laboratory. Additional partners include IBM and Planet. FDL is hosted by the SETI Institute and NASA Ames Research Center.
FDL’s goal is to apply the powerful synergies between physics, simulation and machine learning — many of which are emerging in the commercial sector — to problems important to space exploration and humanity.
Over the past six years, FDL has successfully demonstrated the potential for interdisciplinary AI approaches to tackle challenges in planetary defense, space weather and lunar prospecting. FDL's researchers have helped move the state-of-the-art in using AI to predict solar activity, map lunar resources, build 3D shape models of potentially hazardous asteroids, discover uncategorized meteor showers and determine the efficacy of asteroid mitigation strategies.
FDL tackles knowledge gaps in space science by pairing machine learning experts with researchers in astronomy, astrophysics, astrobiology and planetary science. They work together for an intensive eight-week research sprint, held in the summer break of the academic year - although the journey from challenge definition through to finished result (tech memo and trained algorithm and data products) takes 12 months.
Interdisciplinary four-person teams of PhD and postdoc level researchers and a faculty of three domain and ML experts address tightly defined science challenges informed by knowledge of "what's possible in ML." The faculty, who are subject matter experts, provide support to the teams, drive research quality, and push to more ambitious solutions. External and partner experts, special guests, and reviewers from space agency stakeholders contribute to understanding the problem and providing a community of expertise that drives excellence.
FDL 6.0 will build upon the work, processes and learning developed over the previous five cycles, with the potential to deepen the impact of the work and advance science in new ways.