ASCR FOA: Machine Learning and Understanding for High Performance Computing Scientific Discovery

ASCR has announced a new FOA. Details HERE.

Announcement Number: DE-FOA-0001575
Post Date: 04/18/2016
Close Date: 06/21/2016

  • Letter of Intent Due Date: Not Applicable
  • Pre-Application Due Date: 05/20/2016 at 5 PM Eastern Time
    • (A Pre-Application is required)
  • Encourage/Discourage Date: 05/25/2016 at 5 PM Eastern Time
  • Application Due Date: 06/21/2016 at 11:59 PM Eastern Time


The Office of Advanced Scientific Computing Research (ASCR) in the Office of Science (SC), U.S. Department of Energy (DOE), invites proposals for basic research that significantly advances Machine Learning and Understanding for High Performance Computing Scientific Discovery in the context of emerging algorithms and software for extreme scale computing platforms and next generation networks. The Department of Energy has the responsibility to address the energy, environmental and nuclear security challenges that face our nation. The mission of the Office of Science is the delivery of scientific discoveries and major scientific tools to transform our understanding of nature and to advance the energy, economic, and national security of the United States.

In the exascale computing timeframe, scientific progress will be predicated on our ability to process large, complex datasets from extreme scale simulations, experiments and observational facilities. Even at present, scientific data analysis is becoming a bottleneck in the discovery process; we can only assume that the problem will become more so in the coming decade. At the moment, scientists are often forced to create ad hoc solutions where a lack of scalable analytic capabilities means that there are large-scale experimental and simulation results that cannot be fully and quickly utilized. Moreover, the scientists lack dynamic insight into their analyses, unable to modify the experiment or simulation on the fly. How could we enable broadly applicable solutions to address these challenges?

In this program, we envision that Machine Learning and Understanding may offer the potential 2 to transform basic scientific research best practices, by enabling systems to self-manage, heal and find patterns and provide tools for the discovery of new scientific insights. The goal of this program is to enable and identify basic fundamental research challenges to enable extreme scale machine learning and understanding focusing specifically on high performance computing challenges.