BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//RLASKEY//CALENDEROUS//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VEVENT
DTSTAMP:20260915T200536Z
LAST-MODIFIED:20210423T220449Z
DTSTART:20210429T153000Z
DTEND:20210429T163000Z
UID:event2426@bu.edu
URL:http://physics.bu.edu/internal/events/show/2426
SUMMARY:Forecasting the beta-FPUT Recurrences with Machine Learning Models
DESCRIPTION:Featuring Matthew Litton\, Physics Department\, BA Candidate\n\
	nPart of the Thesis Defenses.\n\nThe Fermi-Pasta-Ulam-Tsingou (FPUT) proble
	m is a cornerstone in the study of nonlinear systems\, statistical mechanic
	s\, and experimental/computational mathematics. We will approach this probl
	em by developing machine learning models to approximate the beta-FPUT dynam
	ics with a focus on correct predictions of future recurrences to initial co
	nditions. The beta-FPUT chain is a high-dimensional nonlinear system which 
	in general proves challenging for machine learning algorithms and so serves
	 as a test of the capabilities of modern deep neural networks. The base arc
	hitecture for our model is a Long Short-Term Memory/Dense Neural Network (L
	STM-DNN) hybrid model. We will further address some modifications to this a
	rchitecture\, primarily with a focus on regularization\, using both model i
	ndependent L2 regularization and a Physics Guided Neural Network model (PGN
	N) which incorporates an unsupervised energy conserving penalty term to the
	 loss function. The models are then evaluated by three metrics\, the root m
	ean squared error (RMSE) between the predicted and the true solution\, recu
	rrence times\, and energy difference.\n\n\n----\n\nTime: Apr 29\, 2021 11:3
	0 AM Eastern Time (US and Canada)\n\nJoin Zoom Meeting\nhttps://bostonu.zoo
	m.us/j/91830715048?pwd=L3FvamRQTmtrcUZqUDlaWnByTFlYUT09\n\nMeeting ID: 918 
	3071 5048\nPasscode: 659173
LOCATION: \, \, 
STATUS:CONFIRMED
CLASS:PUBLIC
END:VEVENT
END:VCALENDAR