White House National Strategic Computing Initiative Workshop Proceedings
Summary of a Workshop, October 20-21, 2015, McLean, VA
The White House National Strategic Computing Initiative Workshop Proceedings is a summary of the workshop held at the Hilton McLean Tysons Corner in McLean, Virginia on October 20-21, 2015. The Networking and Information Technology Research and Development (NITRD) program National Coordination Office (NCO) has edited and released this report in support of the White House National Strategic Computing Initiative (NSCI).
This workshop was made possible by the financial and organizational support of the U.S. Department of Energy on behalf of the NSCI Executive Council.
The second panel was moderated by Dona Crawford (Lawrence Livermore National Laboratory) and featured Chris Johnson (University of Utah), David Bader (Georgia Tech), Peter Koenig (Procter and Gamble), Anna Michalak (Carnegie Institution for Science), and Fred Streitz (Lawrence Livermore National Laboratory) as panelists.
Dona Crawford initiated the discussion by asking how to mitigate risk involved in “bad analytics” or “bad data.” David Bader remarked that traditionally, HPC is concerned more about accuracy, while data analytics is more focused on making decisions, which entails embracing some degree of noise and risk. Michalak commented that in environmental modeling research, all data is imperfect data, as surrogate data is often employed. When informational content is limited, it is tempting to incorporate all that was collected – the central challenge is deciding what should be excluded from the analysis.