NOTE NEW TIME AND LOCATION

Abstract: Materials science is undergoing profound changes, driven by advanced machine learning algorithms, continual improvements to instrumentation that have resulted in an explosion in the data volume, dimensionality, complexity, and variety, and increased access to high performance computing resources. This talk will focus on the software, tools, and infrastructure efforts at Oak Ridge National Laboratory (ORNL) that aim to facilitate data-driven scientific discovery in this era of big data.

Many of these technologies could be important components of a successful laboratory information management system. Starting from data acquisition, we have developed a data schema capable of representing observational data of any size, dimensionality, and modality. Data structured in this manner and written into open file formats overcomes the problems of multiple proprietary file formats, and development of generalized data analysis algorithms. Such data are analyzed and visualized using a family of community-driven software packages, especially Pycroscopy, that apply cutting-edge machine learning techniques to materials science.

This talk will also discuss ongoing efforts on developing software, infrastructure, and policies for integrating experimental facilities to computational and data facilities for leveraging high-performance computing for analysis. The talk will also discuss other development of software that facilitates seamless data management, searching, sharing and tools for robust transfers of very large data files across the globe. In an effort to facilitate data-driven scientific discoveries, efforts are underway to deploy a data catalog for gathering, curating, and using scientifically important datasets in addition to automatically granting unique digital object identifiers for each dataset in the catalog.

Bio: Suhas Somnath is a computer scientist at the National Center for Computational Sciences at ORNL, straddling the physical and computational domains in finding artificial intelligence, computing, and infrastructure solutions for problems in the domain sciences.