POSTPONED: HEMI Colloquium: Mike Shields

In the interest of building community and collaboration in HEMI, we are initiating the HEMI Friday Colloquium series. Speakers in this series will be members of the JHU community, both from within and from outside of HEMI. They will occur about once per month, initiating with a social time from 3:00-3:30 with snacks and beverages, followed by a talk from 3:30-4:30, with snacks and beverages continuing. The idea is for this to be much more casual than the typical seminar, with a lot of discussion and interaction.

This event has been postponed to a later date. Please check the calendar for the rescheduled date and time.

Stochastic simulation-based uncertainty quantification in computational physical modeling

Computational physical modeling has undoubtedly changed the landscape of analysis and design across numerous disciplines. With rapid advances in deterministic numerical modeling tools, our ability to model increasingly complex multi-physical and multi-scale phenomena is constantly improving. Yet, the ability to accurately model individual scenarios under tightly-controlled conditions often does not translate to robust and accurate predictions of the physical world given the abundance of uncertainties associated with either inherent variability of physical phenomena or our lack of knowledge/data describing our system or its environment. To extend our physical models to practice requires a careful and rigorous accounting and propagation of these many uncertainties. In this presentation, a survey of new stochastic simulation-based methods developed by the Shields Uncertainty Research Group (SURG) at Johns Hopkins University are presented that enable end-to-end uncertainty quantification in computational physical modeling and are computationally tractable for large, complex, nonlinear, and dynamic systems. Emphasis is placed on demonstrating their abilities to improve structural safety and reliability.

Seminar will be held in Malone Hall, Room G33/35.