Scientific Discovery from Chemical Data Analyses

University Of Delaware, Newark DE

Investigators

Abstract

With support from the Division of Chemistry and partial co-funding from the Divisions of Mathematical Sciences and the Office of Advanced Cyberinfrastructure, Professors Karl Booksh and Sharon Neal at the University of Delaware are organizing a workshop to identify new frontiers in data analyses associated with chemical measurements for scientific discovery. Advances in scientific instrumentation have helped enable the "data revolution" through collection of vast amounts of data, both to probe a single sample in exquisite detail and to collect snippets of data from a wide array of interesting samples. Data analytic strategies, often under the banner of machine learning and artificial intelligence for example, offer new tools to help researchers gain valuable scientific insights from such data collections. Unfortunately, there remains an adaptation gap between these tools and researchers with application domain knowledge who would benefit from the availability of more advanced data analysis methods. This workshop seeks to identify convergent areas of chemical research that would most benefit from new cutting-edge tools for data analysis and to identify means of reducing the impediments to broad and rapid adoption of such tools by researchers with relevant application domain knowledge. The workshop will also explore needs and means for training of students in valuable data analysis skills. The workshop will bring together researchers in the chemical sciences and related areas (e.g., environmental analysis, biological chemistry, materials science, and national security monitoring) and experts in novel data analytics (e.g., machine learning, artificial intelligence, graph theory). Participants will explore the relationship between pressing scientific questions in convergent application fields and the potential for new analytical methods that uniquely address these questions. For example, how might graph theory or network analyses advance prediction of new catalysts and more efficient reaction sequences? What insights might machine learning or artificial intelligence derive from high-resolution, multivariate images of catalytic surfaces? How can sensor arrays be best adapted to improve understanding and prediction of the chemistry of dynamic environmental processes? Participants will strive to identify where collaborations are necessary to advance convergence research and means of facilitating establishment of those collaborations. The workshop will also consider the education and training needs for creating a research workforce versed in both measurement science and data analysis. The workshop will engage approximately 30 thought-leaders spanning academic, industrial, and governmental sectors using virtual and/or in-person formats. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.