This two-year CyberTraining pilot project at Clarkson University will conduct a two-week online summer workshop each year for six trainees in engineering and science related disciplines as well as integrate the same workshop material into an existing graduate-level course on High-performance Computing (HPC) to broaden its appeal to graduate students in engineering and science. As shown in Fig. 1, the crux of our research workforce development program is on POD, a highly effective, data-driven learning algorithm for solving multi-dimensional PDEs. Our workshop trainees will be graduate students, post-docs and faculty members recruited nationally. We propose to train three teams each year of two trainees per team in the summer workshop, where each trainee team will be supported by a research mentor (one of the three PIs) and a graduate teaching assistant (TA).
This CyberTraining pilot program provides development of the national research workforce in areas of critical need through (i) intensive, integrated instruction on using open-source computing platforms to solve Ordinary/Partial Differential Equations (ODEs/PDEs), and develop Proper Orthogonal Decomposition (POD) models using HPC, (ii) research training in team-based interdisciplinary projects that offer an effective approach based on a datadriven leaning algorithm (POD) for computationally intensive multiphysics simulation problems in various science and engineering disciplines, and (iii) piloting and integration of the instructional material developed from the summer workshop into an existing graduate course on HPC. Unfortunately, POD is rarely covered in the current graduate curriculum and thus the nation is not fully leveraging this advanced algorithm in science and engineering research yet. Our proposed summer workshop and the integration of the same material into an existing graduate course provide our trainees and students with an intensive but streamlined instruction on POD and related prerequisite topics (e.g., utilizing open source platforms to solve PDEs and eigenvalue problems). The training culminates in a week-long research project where trainees and students learn advanced computational tools and implementation skills (PETSc [1][2] , SLEPc [3], FEniCS [4], HPC) in a hands-on manner while enjoying the close guidance from the PI team and a dedicated TA. Through these proposed activities, we target both solicitation goals of broadening adoption of advanced CI (i.e., using open-source platforms to solve ODEs/PDEs and implement POD using HPC) and integration of the CI skills into the existing graduate curriculum/instructional material fabric.