SEAS trainee, Jocelyn Chi, participated in the Oak Ridge Institute Science and Education (ORISE) National Science Foundation (NSF) Mathematical Sciences Graduate Internship (MSGI) program at Lawrence Berkeley National Laboratory. She interned with Dr. Xiaoye Sherry Li in the Scalable Solvers Group of the Computational Research Division. Jocelyn was among over 250 interns in the Computational Research Division, and 11 in the Scalable Solvers group, which consisted of faculty, undergraduates, and doctoral students with backgrounds in computer science, physics, or math. She was the only statistician. During the 10-week immersion, Jocelyn worked on the project, A Scalable Algorithm for Approximate Generalized Eigenspaces of Fock Hamiltonian Matrices, which converges at the intersection of numerical linear algebra, high performance computing, and computational quantum chemistry. More specifically, she investigated how to utilize matrix approximations and their corresponding factorizations and linear algebra routines – in the Shift-Invert method to produce a scalable algorithm for high-fidelity approximate solutions of the Hartree-Fock equations derived from quantum chemistry data.
Jocelyn had the opportunity to attend a number of seminars throughout the summer, and participate in a number of workshops including the National Energy Research Scientific Computing (NERSC) New User Training workshop on how to use the high-performance supercomputers at NERSC, a two-day workshop entitled “Crash Course on Supercomputing,” and the LBL Molecular Foundry and Computational Sciences Area Summit designed to facilitate discussions and interactions between researchers at the cutting edge of technological advancements in materials research at the Molecular Foundry, computational mathematicians in the Computational Research Division, and high performance computing specialists at NERSC. Jocelyn also participated in the Institute for Computational and Experimental Research in Mathematics (ICERM) Women in Data Science and Mathematics (WiSDM), a week-long immersive program to convene women at various stages to collaborate on defined data science problems in Providence, Rhode Island. Through this internship experience, Jocelyn gained a working knowledge of C++, acquired skills for high performance computing, and hands-on introduction to eigenvalue problems and iterative algorithms.
Jocelyn’s reflects, “My immersion in Sherry’s research group this summer was a richly rewarding experience. I really appreciated the collaborative environment and learned so much from discussions with many of the members of the Scalable Solvers Group at LBL. In addition to the many things I learned and skills I developed this summer, I worked in a research area that I was quite unfamiliar with, and developed confidence over the weeks in my ability as a researcher to pick up new skills and knowledge with sustained effort. Moreover, after seeing the kinds of problems that people can work on at the intersection of mathematical and computational research, I’ve gained a lot of clarity about the kind of research career I’d like to have, and the effort and training that I’ll need to invest towards that goal.”
For more information on the MSGI program, visit: https://orise.orau.gov/nsf-msgi/