DOE to provide $30M for new data science approaches for energy-relevant chemistry and materials research
Data science combines computer science, applied mathematics, and statistics with domain science to discover new knowledge from often complex (such as unstructured) data sets generated from experimental and/or computational studies.
As part of data science, machine learning and artificial intelligence methods are rapidly evolving and leading to more accurate predictions and trustworthy decisions and actions. While scientific research has benefited greatly in many areas such as the chemical and materials sciences as well as bioinformatics, medicine, drug discovery, systems control, astronomy and particle physics, many opportunities remain for data science to accelerate the rate of fundamental discovery for complex chemical processes and materials, the DOE said.
Synergistically complementing theory, experimentation, and simulation, data science has already demonstrated some success in accelerating molecular and materials discovery. The next goal for this methodology is to generate new fundamental understanding of physical and chemical behavior, leading to successful predictions well outside the range of the original data, and possibly forcing modifications to the theoretical approaches.
One specific area of interest to BES-supported research is that of complex chemical and materials processes, defined as those whose macroscopic properties, reaction mechanisms and dynamic behavior cannot be predicted from a known combination of the properties of the individual components, hence are outside the scope of existing theoretical approaches.
Complex systems involve massive combinatorial spaces and nonlinear processes, which are being slowly tackled with large experimental, theoretical and computational efforts. Data science approaches, in combination with standard experimental and theoretical methods, are expected to accelerate the discovery of fundamentally new chemical mechanisms and material systems with exceptional properties and dynamic behavior, as well as new physical principles or laws.—DE-FOA-0002082
The initiative seeks proposals that focus on innovative applications of modern data science approaches to understand processes and mechanisms in complex energy-relevant chemical and materials systems.
National laboratories, universities, nonprofits, and companies will be eligible to apply for the three-year awards, which will be selected on the basis of peer review. The Basic Energy Sciences program in the Department’s Office of Science, which is funding the effort, envisions awards for single investigators and small groups.
Pre-applications are due on 8 March 2019; the deadline for final applications will be 15 May 2019. Planned funding for Fiscal Year 2019 will be $10 million, with outyear funding contingent on congressional appropriations.