Other

Machine Learning

Finally succeeding in my effort to make my undergrad linear algebra classes worthwhile, I’ve taken classes and done some side projects in machine learning. My (astro-related!) CS229 final project compared several ML methods performance at star-galaxy separation, the critical task of classifying all the astronomical sources measured by a large survey as being stars (slightly smaller and bluer) or galaxies (slightly larger and redder). You can read my final report here.

I took on a different domain for CS231n, the Stanford course on convolutional neural networks–art history! Using a modified VGGNet architecture, we were able to achieve state-of-the-art+ performance on identifying paintings by artist, nationality, and genre! Our report (including some trippy class-optimized images) can be found here.

Outreach

I’ve taken part in all sorts of fun outreach activities while at Stanford: giving tours of our visualization lab at the KIPAC Open House, being quizzed at “Quiz-a-Cosmologist” at the College of San Mateo’s Astronomy fair, and answering public questions over drinks at astronomy-themed “Nightlife” events at the California Academy of Sciences. I’ve also taken part in the Science in Service program through Stanford’s Haas Center for Public Service, where we workshopped ways to improve K-8 outreach and put them into practice as tutors at the after-school science program at the Boys and Girls’ Club of Redwood City, CA.

DC Trip

Last year, and again in March 2017, I travel to Washington, D.C. with a group of scientists from Fermilab, SLAC, and the US LHC Users’ Organization to spread the message to Congressional and Executive Branch staff that physics research contributes to the economy and to American education by inspiring future generations of STEM students. Highlights from last year included a visit to Pat Dehmer at DOE HQ on the National Mall, and mesmerizing Congressional staffers with the rulers made from wavelength-shifting crystal we brought along!