Machine Learning is a field of computer science that gives computers the ability to learn without being explicitly programmed. Machine learning algorithms enable computers to make predictions or recommendations based on data, patterns, and experiences. Learning-capable machines can already forecast next-day demand or network failure on select electric grids. They can also pick out individual signals from apparent noise, much like a human can identify two different songs playing concurrently on a busy street. Yet, today’s best computer processors cannot process all the information surrounding an autonomous vehicle, and the energy system is even more data intense. This forum provides cross sections of what is possible—with implications for subsurface mapping, material development, preventive maintenance, and power plant cybersecurity. Presenters: Mr. Prabhat, NERSC, and Dr. George Guthrie, LANL
Presentations
Machine Learning for Science (Mr. Prabhat)
Machine Learning Opportunities (Dr. George Guthrie)
Video
Biographies
Prabhat
Prabhat leads the Data and Analytics Services team at NERSC. He is responsible for the Data stack (spanning Analytics, Management, Workflows, Visualization, and Access/Transfer tools) on NERSC platforms. Prabhat is also the Director of the Intel and Cray-sponsored Big Data Center, which is enabling capability applications on the Cori supercomputer.
Prabhat has contributed to over 100 publications spanning topics in Computer Science (Big Data, Statistics, Machine Learning, Data Management, HPC, Scientific Visualization) and Domain Sciences (Climate, Planetary Geoscience, Physics, Chemistry, Neuroscience). Prabhat received an ScM in Computer Science from Brown University, a B. Tech in Computer Science and Engineering from IIT Delhi, and is currently pursuing a PhD in Earth and Planetary Science from UC Berkeley.
George Guthrie
Dr. George Guthrie is a mineralogist/geochemist at Los Alamos National Laboratory. He received his AB from Harvard University (1984) and PhD from Johns Hopkins University (1989), after which he was a postdoctoral fellow and subsequently staff member, deputy group leader, program manager, and program director at Los Alamos. In 2008 he became an employee of the Office of Fossil Energy in the US Department of Energy when he joined Office of Research and Development at the National Energy Technology Laboratory.
In 2014, he re-joined Los Alamos in the Earth and Environmental Sciences division, where he also supports the Office of Applied Energy Programs (with responsibility for LANL’s fossil-energy and geothermal-energy portfolios). His research areas include a variety of energy-related topics, particularly related to subsurface systems, including CO2 sequestration, cement integrity, shale gas, and risk assessment. In this latter area, while at NETL he initiated the National Risk Assessment Partnership (NRAP), for which he served as technical director and now chair of the executive committee; NRAP has led development of a new approach to predicting the behavior of complex subsurface systems that combines physics-based prediction in combination with empirical models (including machine learning).