Had the opportunity to come out and visit the Army Research Lab this December for a 2-week visit to the ARL to see their facilities and get to know their researchers. This was a hip-fire opportunity, basically I sat down with a couple folks from the ARL on a conference call and they were interested in the materials research, as well as my experience of being an officer in the Army. They invited me out to ARL in Aberdeen, and were able to pay for it through DTS, which only worked because I'm in the Army too, by happenstance.
We had the opportunity to sit down and work with some of their equipment, as well as work with some of their researchers and visit with them about what they were working on and how we could collaborate. They have an incredible amount of data, mostly the directorate is experimental, except for the small computational group I was working with. This provided some awesome opportunity to sit down with some experimentalists in the materials space and pick their brains on some projects we are working on. They were a little less than receptive on some of the microstructure reconstruction tasks I was working on, but thats alright.
It was great to get to know them, and work with them. We discussed me coming out for the summer, which I think would be great.
Update, March 2019: Dr. Haile has agreed to bring me out this summer, so I'll be headed out there and getting a little studio apartment for the summer and working in MD!
Interned this summer at Army Research Lab in Aberdeen, MD under the supervision of Dr. Mulugeta Haile.
Interestingly enough, no application was required for this, we started working together in the Fall and now here we are. They have a special program to bring out researchers from universities, and PhD students are included in that. So I get the opportunity to work with Haile and his computational team out here, which has been great.
I have been working the GAN project I started in MSEN 655 (see Courses reflections for that post) and have had some pretty promising results. We have been able to generate some really high quality microstructure results:
Here we see a subset of the microstructures. These were generated using a progressively trained GAN [1], trained using a DGX style server with 8 Tesla V100 GPUs.