Just a few of the 92 students who participated in research at Argonne’s Mathematics and Computer Science division this summer
This past summer more than 90 students joined the MCS division to work with staff members tackling problems in computational mathematics and computer science. The students ranged from undergraduates to doctoral candidates and came from universities throughout the world (see the maps with a pinpoint on the academic institutions – change to throughout the nation if you use the U.S. map).
The student researchers were supported by a variety of programs, including DOE’s Computational Science Graduate Fellowship (CSGF) program, NSF’s Mathematical Sciences Graduate Internship (MSGI), DOE’s Science Undergraduate Laboratory Internship (SULI), the University of Chicago Metcalf fellow program, Argonne’s Student Research Participation program, and the MCS Division’s named Givens and Cody Associates programs. Here we give just a few examples of the exciting projects MCS interns worked on.
For example, Tiago Correia spent three months at Argonne this summer through the NSF MSGI program, researching scalable parallel optimization algorithms. A graduate student at Columbia University, Correia said he is “interested in all things mathematical optimization, data science, and numerical methods.” At Argonne he worked with MCS division computational mathematicians on revisiting classical optimization methods for increasing the scope for massive levels of parallelism.
Huansong Fu is a Ph.D. candidate in computer science at Florida State University. He worked with the programming models and runtime systems group in the MCS division as a research aide during the summer. His research involved analysis and optimization of an OpenSHMEM runtime on the Message Passing Interfaee.
Adam Morrison joined Argonne this summer supported by the Science Undergraduate Laboratory Internship (SULI program). SULI encourages undergraduate students to pursue science, technology, engineering, and mathematics careers by providing research experiences at the DOE laboratories. Morrison, a computer engineering student at Valparaiso University, says he is passionate about pushing the envelope in what technology can do and in working on the integration of hardware and software. He had a chance to do just that, working in both 2017 and 2018 with MCS staff on the Waggle intelligent sensors project.
Nimish Awalgaonkar, a graduate student in the School of Mechanical Engineering at Purdue University, worked in the MCS division as a Givens Associate. These positions, named after noted mathematician Wallace Givens, are intended to encourage doctoral students in numerical analysis or computational mathematics. Awalgoankar said, “My lab director at Purdue had been a postdoctoral researcher in the MCS division, and he encouraged me to apply for the Givens Associate fellowship. At MCS, I was able to explore Bayesian optimization methods using deep neural networks. Neural networks are a new programming paradigm inspired by biology, and deep learning is a branch of machine learning that scientists are exploring in combination with neural networks as a way to solve problems scientific machine learning.”
Morgan Kelley is working on her Ph.D. in ways to improve the efficiency of power grids and save energy. The recipient of a DOE CSGF award, she joined the MCS division to expand her research to develop production scheduling strategies and algorithms to mitigate fluctuations in demand on the power grid. CSGF fellows are provided up to four years of financial support for pursuing doctoral degrees in areas that use high-performance computing to solve complex problems in science and engineering.
Tucker Hartland, an applied mathematics graduate student at the University of California, Merced, spent the summer working on PETSc, the Portable, Extensible Toolkit for Scientific computing. PETSc is a suite of libraries developed at Argonne for solving scientific applications modeled by partial differential equations (PDEs) and includes the Toolkit for Advanced Optimization. Hartland investigated PDE-constrained optimization in PETSc for elliptic, parabolic, nonlinear, and time-dependent problems. He worked in the MCS division as a Givens Associate, and he stated, “The experience was enriching to my continuing development as an applied mathematician.”
Not all the summer interns at MCS were involved in developing algorithms or software. Priya Linguitla worked as a science writing intern. An undergraduate at the University of Chicago, Linguitla applied for the internship because she wanted to learn how to effectively communicate complex research to the public. Her writing activities at Argonne included interviewing staff members on topics ranging from message-passing programming models and data movement to image processing and wireless sensors. “Working with my team of science writers and communicators was truly an incredible experience, and getting to know what drives the MCS researchers — their ambitions and interests alike — has been the highlight of my summer,” Lingutla said.
Students presented their work in a variety of venues, including multipart MCS student seminars and the LANS Summer Argonne Student Symposium (SASSy) sessions. “It was great to learn about the broad array of research accomplished by students this summer; having students distill their summer’s work in a concise talk is a great experience and provided an exciting exposition of recent work in MCS,” said Matt Menickelly, a chair of one of the SASSy sessions. Currently a postdoc in MCS, Menickelly knows the experience well, having himself been a Givens Associate during his graduate studies.