BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//RLASKEY//CALENDEROUS//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VEVENT
DTSTAMP:20261019T165304Z
LAST-MODIFIED:20180904T160845Z
DTSTART:20180906T193000Z
DTEND:20180906T203000Z
UID:event2012@bu.edu
URL:http://physics.bu.edu/internal/events/show/2012
SUMMARY:Machine learning and glassy optimization of the quantum control pro
	blem
DESCRIPTION:Featuring Alexandre Day\, Physics PhD Candidate at Boston Unive
	rsity\n\nPart of the Departmental Seminars.\n\nWe study the problem of prep
	aring a quantum many-body system from an initial to a target state by optim
	izing the fidelity over the family of the so-called bang-bang protocols. We
	 present compelling numerical evidence for a universal spin-glass-like tran
	sition controlled by the protocol time duration. The glassy critical point 
	is marked by the occurrence of an extensive number of protocols with close-
	to-optimal fidelity and with a true optimum that appears exponentially diff
	icult to locate. Using a machine learning (ML) inspired framework based on 
	the manifold learning algorithm t-SNE\, we are able to visualize the geomet
	ry of the high-dimensional control landscape in an effective low-dimensiona
	l representation. Across the glassy transition\, the control landscape feat
	ures a proliferation of an exponential number of attractors separated by ex
	tensive barriers\, which bears a strong resemblance with replica symmetry b
	reaking in spin glasses and random satisfiability problems. We further show
	 that the quantum control landscape maps onto a disorder-free classical Isi
	ng model with frustrated nonlocal\, multibody interactions. Our work highli
	ghts an intricate but unexpected connection between optimal quantum control
	 and spin glass physics\, and how tools from ML can be used to visualize an
	d understand glassy optimization landscapes.
LOCATION:SCI 328\, 590 Commonwealth Avenue\, 02215
STATUS:CONFIRMED
CLASS:PUBLIC
END:VEVENT
END:VCALENDAR