Student Contributions

David Brough has just released a latest version of PyMKS, an open access, open source, code repository aimed at hierarchical material systems. PyMKS 0.2 adds functionality for microstructure quantification through high level APIs. These functions are used to create homogenization linkages and complement the localization linkage code base that was present in PyMKS 0.1. In addition to the new functionality, PyMKS 0.2 will used PyFFTW if installed to boost performance. An updated website with two new examples were added to demonstrate the new functionality as well as additional documentation (http://pymks.org).

Perry Ellis has been exposed to various computer vision methods that have proved crucial to extracting data from our experiments.  These relay on large amounts of 3D confocal microscopy images.  Perry has learned crucial statisticalmethods and ideas that have been essential in our experiments. 

Jason Allen won the 2016 MSE Outstanding TA of the Year and the NSF Graduate Research Fellowship Program Honorable Mention.

Christopher Shartrand has successfully created a script of CV code that automatically extracts physical features from AFM images of organic field effect transistors. This is significant as the current baseline methods involve humans using computer applications to painstakingly extract the features themselves. This script of CV code is fundamental to the structure property modeling that my team is currently working on.

Robert Pienta successfully proposed his PhD thesis; Peter Griffiths and Andriy Dotsenko both passed their PhD Qualifying exams.

Christopher Shartrand aided in writing and editing a large scale NSF proposal under the direction of Drs. Martha Grover, Elsa Reichmanis, and JC Lu.  If the grant is awarded, it will significantly aid the future of the research that is being conducted in process-structure-property relationships of organic field effect transistors.

Robert Pienta has learned a tremendous amount about the challenges and opportunities in different domain areas.  Specifically that there may be some low-hanging fruit for mapping PSP linkages in metallurgy.  He is in the early stages of collaboration with Ashley Goulding and members of her lab.  Robert is also working with Almambet Iskakov and has successfully modeled the time-varying behavior of a family of ternary eutectoid alloys.  He has also began working with David Brough on using privileged learning models to improve Process-Structure linkage performance.

Alex Lohse received an NSF Graduate Research Fellowship Program Honorable Mention and was a Center for the Enhancement of Teaching and Learning TA Award Finalist.  Alex also placed 3rd in the Polymers Category at the Annual MSE Poster Competition.  Alex was able to use local structure space information in a molecular dynamics simulation of covalent adaptable networks as features in various machine learning algorithms to correctly predict bonding conditions. This is important because it allows us to build a model that when re-introduced as a routine in the simulation will essentially cut the simulation time in half compared to the old way we were simulating these materials. Separately, he was able to use spatial correlations and PCA to classify and characterize interfacial epoxy structure in graphene/epoxy composites. Previously in molecular dynamics there was no definitive way to classify structure (especially epoxy since it is inherently amorphous). This was a way to prove that graphene functionalizations change the interfacial structure of epoxy. The goal is now to use this information to tailor the structure to give the materials the properties we desire.