BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Physics-informed machine learning - Pingfang Song (University of C
 ambridge)
DTSTART:20230329T100000Z
DTEND:20230329T113000Z
UID:TALK199063@talks.cam.ac.uk
CONTACT:James Allingham
DESCRIPTION:This talk will introduce some of the prevailing trends in embe
 dding physics into machine learning.\n\nMachine learning has shown great p
 otential in tackling challenging tasks in a variety of fields. However\, t
 raining deep neural networks requires big data\, not always available for 
 scientific problems. Instead\, one alternative approach is to utilize addi
 tional information based on the laws of physics to train these networks. S
 uch physics-informed learning integrates (noisy) data and mathematical mod
 els\, and implements them through neural networks or other kernel-based re
 gression networks. Moreover\, it may be feasible to design customized netw
 ork architectures that inherently satisfy physical constraints\, leading t
 o enhanced accuracy\, faster training and improved generalization.\n\nRefe
 rences (recommended reading\, not required):\n\nKarniadakis\, George Em\, 
 et al. "Physics-informed machine learning." Nature Reviews Physics 3.6 (20
 21): 422-440.\n\nLu\, Lu\, et al.. "DeepXDE: A deep learning library for s
 olving differential equations." SIAM review\, 2021.\n\nRaissi\, Maziar\, P
 aris Perdikaris\, and George E. Karniadakis. "Physics-informed neural netw
 orks: A deep learning framework for solving forward and inverse problems i
 nvolving nonlinear partial differential equations." Journal of Computation
 al physics 378 (2019): 686-707.\n\nLu\, Lu\, et al. "Learning nonlinear op
 erators via DeepONet based on the universal approximation theorem of opera
 tors." Nature machine intelligence 3.3 (2021): 218-229.\n\nKovachki\, Niko
 la\, et al. "Neural operator: Learning maps between function spaces." arXi
 v preprint arXiv:2108.08481 (2021).\n\nLi\, Zongyi\, et al. "Fourier neura
 l operator for parametric partial differential equations." arXiv preprint 
 arXiv:2010.08895 (2020)
LOCATION:Cambridge University Engineering Department\, CBL Seminar room BE
 4-38.
END:VEVENT
END:VCALENDAR