BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
X-WR-CALNAME:Distinguished Seminar in Computational Science and Engineering
X-WR-TIMEZONE:Eastern Time (US & Canada)
BEGIN:VEVENT
DTSTAMP:20261106T140035Z
UID:tag:localist.com\,2008:EventInstance_51950776359321
DTSTART:20260305T170000Z
DTEND:20260305T180000Z
DESCRIPTION:Distinguished Seminar in Computational Science and Engineering\
 n\n \n\nMarch 5\, 2026\, 12-1PM\n\n \n\n45-432 in Building 45 and Zoom Web
 inar\n\n \n\nLeveraging Machine Learning for Simulations of Turbulent Mult
 i-Physics Flows: Computational Acceleration\, Physics Generalization\, and
  Model Hybridization\n\n \n\nMichael Mueller\nDonald R. Dixon ’69 and El
 izabeth W. Dixon Professor\nMechanical and Aerospace Engineering\nPrinceto
 n University\n\n \n\nAbstract: \n\nIntegrating machine learning algorithms
  into computational simulations of turbulent multi-physics flows has the p
 otential to improve these simulations in every sense of the word. Three di
 stinct use cases for machine learning within computational simulations are
  presented and example applications drawn from combustion\, plasmas\, and 
 offshore wind. First\, machine learning algorithms can be used to accelera
 te computational simulations by reducing redundant computations. In solvin
 g a set of governing PDEs\, one often must repeatedly evaluate algebraic a
 uxiliary relationships (e.g.\, equations of state\, constitutive laws\, mo
 dels for unresolved physics\, etc.). These repetitive evaluations limit th
 e computational intensity of these auxiliary relationships and/or the numb
 er of inputs. Adaptive binary trees are leveraged to learn these auxiliary
  relationships ‘on-the-fly’ and reuse prior evaluations with strict er
 ror control. Second\, machine learning algorithms can be used to generaliz
 e these auxiliary relationships\, with an emphasis on models for unresolve
 d physics. These adaptive binary trees enable the use of more general auxi
 liary relationships that might otherwise be too computationally or memory 
 intensive. However\, as the complexity increases\, the binary trees can th
 emselves become too memory intensive. One advantage of machine learning al
 gorithms such as neural networks is their compact storage\, albeit at the 
 cost of often considerable training time. A neural adaptive binary tree ap
 proach is developed in which regions of the adaptive binary tree are perio
 dically pruned and the constituent data used to train a shallow neural net
 work to replace this region of the tree\, substantially reducing memory re
 quirements. Strategies have also been developed for error control. Third\,
  machine learning algorithms can be used to develop hybridized physics-bas
 ed and data-based models. Rather than an either/or paradigm\, hybridizatio
 n leverages a both/and paradigm. While physics-based modeling frameworks a
 re preferred due to their explicit satisfaction of fundamental physical la
 ws\, physics-based approaches often rely on sometimes invalid assumptions 
 with no clear pathway forward (e.g.\, closure modeling). In a weak sense\,
  data-based approaches can be used to provide new insights for missing phy
 sics and physics-based model improvement. In a strong sense\, a data-based
  model can be trained to replace the hardest-to-model terms in an otherwis
 e physics-based framework\, retaining its attractive physical features.\n\
 n \n\nBio:\n\nMichael E. Mueller is the Acting Chair and Donald R. Dixon 
 ’69 and Elizabeth W. Dixon Professor of Mechanical and Aerospace Enginee
 ring and the Interim Director of the Princeton Institute for Computational
  Science and Engineering at Princeton University. Since 2020\, he is also 
 jointly appointed as a Faculty Researcher at the National Laboratory of th
 e Rockies. He currently serves as Editor of the Proceedings of The Combust
 ion Institute. He received a BS degree in mechanical engineering from The 
 University of Texas at Austin in 2007\, a MS degree in mechanical engineer
 ing from Stanford University in 2009\, and a PhD degree in mechanical engi
 neering from Stanford University in 2012. His research interests encompass
  computational modeling of multi-physics turbulent flows with applications
  to energy\, propulsion\, and the environment\, including combustion\, off
 shore wind\, aerosols\, and plasmas\, as well as broader areas of computat
 ional and data sciences including uncertainty quantification\, algorithms 
 for heterogeneous computational architectures\, and data-based modeling an
 d ML/AI algorithms. He is a Fellow of the American Society of Mechanical E
 ngineers and an Associate Fellow of the American Institute of Aeronautics 
 and Astronautics. Among other awards and recognitions for his research\, h
 e has been recognized with the Hiroshi Tsuji Early Career Research Award f
 rom The Combustion Institution and the Young Investigator Program (YIP) Aw
 ard from the Army Research Office. He has also received the Princeton Univ
 ersity Graduate Mentoring Award and the School of Engineering and Applied 
 Science Distinguished Teacher Award.
GEO:42.361613;-71.092293
LOCATION:Building 45 (MIT Stephen A. Schwarzman College of Computing)\, 45-
 432
SUMMARY:Distinguished Seminar in Computational Science and Engineering
URL;VALUE=URI:https://calendar.mit.edu/event/distinguished-seminar-in-compu
 tational-science-and-engineering-4518
CATEGORIES:Conferences/Seminars/Lectures
END:VEVENT
END:VCALENDAR
