As more and larger resources of materials data are becoming available, employing machine learning to extract knowledge and predictive models from this information has become a promising path for materials design. While the advantages of machine learning models (e.g., evaluation speed) have been demonstrated previously, the process of constructing new models from materials data is complicated by the lack of established approaches and tools. In order to enable the wider-scale use of machine learning methods in materials engineering, I have developed general-purpose techniques for creating machine learning models from materials data and open-source software capable of being applied to a wide variety of materials problems. In this seminar, I will introduce these tools and demonstrate their application to several materials applications, including the optimization of commercial Bulk Metallic Glass alloys.

Bio:

Logan Ward is a PhD candidate in Materials Science and Engineering at Northwestern University. He earned his BS and MS in Materials Science and Engineering with a minor in Computer and Information Science from The Ohio State University in 2011 and 2012, respectively. His Master’s thesis, completed under the supervision of Wolfgang Windl and Katharine Flores, was on the virtual design of metallic glass alloys using molecular dynamic simulations. In 2012, Logan joined the research group of Christopher Wolverton at NU to study the use of machine learning for engineering and discovering new materials. While a graduate student, Logan has been recognized with a National Defense Science and Engineering Graduate (NDSEG) Fellowship and the Weertman Fellowship from Northwestern University.