BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//jEvents 2.0 for Joomla//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VTIMEZONE
TZID:America/Chicago
X-LIC-LOCATION:America/Chicago
BEGIN:DAYLIGHT
TZOFFSETFROM:-0600
TZOFFSETTO:-0500
TZNAME:CDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0500
TZOFFSETTO:-0600
TZNAME:CST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
UID:ccef1302e9b480425f1ebd8ffdbabd1751
CATEGORIES:Events
SUMMARY:USACM UQ Virtual Seminar
DESCRIPTION:USACM UQ Virtual SeminarLeveraging Physics-Induced Bias in Scientific Machi
ne Learning for Computational Mechanics\n– Physics-Informed, Structure-Pres
erved Learning for Problems with Complex GeometriesSpeakerJian-xun Wang, Un
iversity of Notre Dame\nDiscussant\nMichael Brenner, Harvard University \nA
bstract\n\nFirst-principle modeling and simulation of complex systems based
on partial differential equations (PDEs) and numerical discretization have
been developed for decades and achieved great success. Nonetheless, tradit
ional numerical solvers face significant challenges in many practical scena
rios, e.g., inverse problems, uncertainty quantification, design, and optim
izations. Moreover, for complex systems, the governing equations might not
be fully known due to a lack of complete understanding of the underlying ph
ysics, for which a first-principled numerical solver cannot be built. Recen
t advances in data science. and machine learning, combined with the ever-in
creasing availability of high-fidelity simulation and measurement data, ope
n up new opportunities for developing data-enabled computational mechanics
models. Although the state-of-the-art machine/deep learning techniques hold
great promise, there are still many challenges - e.g., requirement of “big
data”, the challenge in generalizability/extrapolibity, lack of interpreta
bility/explainability, etc. On the other hand, there is often a richness of
prior knowledge of the systems, including physical laws and phenomenologic
al principles, which can be leveraged in this regard. Thus, there is an urg
ent need for fundamentally new and transformative machine learning techniqu
es, closely grounded in physics, to address the aforementioned challenges i
n computational mechanics problems.\n\n\nThis talk will briefly discuss our
recent developments of scientific machine learning for computational mecha
nics, focusing on several different aspects of how to bake physics-induced
bias into machine/deep learning models for data-enabled predictive modeling
. Specifically, the following topics will be covered: (1) PDE-structure pre
served deep learning, where the neural network architectures are built by p
reserving mathematical structures of the (partially) known governing physic
s for predicting spatiotemporal dynamics, (2) physics-informed geometric de
ep learning for predictive modeling involving complex geometries and irregu
lar domains.\nBio\n\nDr. Jian-xun Wang is an assistant professor of Aerospa
ce and Mechanical Engineering at the University of Notre Dame. He received
a Ph.D. in Aerospace Engineering from Virginia Tech in 2017 and was a postd
octoral scholar at UC Berkeley before joining Notre Dame in 2018. He is a r
ecipient of the 2021 NSF CAREER Award. His research focuses on scientific m
achine learning, data-enabled computational modeling, Bayesian data assimil
ation, and uncertainty quantification.\nATTEND
DTSTAMP:20221209T133914
DTSTART;TZID=America/Chicago:20220512T140000
DTEND;TZID=America/Chicago:20220512T150000
SEQUENCE:0
TRANSP:OPAQUE
END:VEVENT
END:VCALENDAR