BEGIN:VCALENDAR
VERSION:2.0
METHOD:PUBLISH
PRODID:-//209.200.118.155//NONSGML iCalcreator 2.2.8//
BEGIN:VEVENT
DTSTAMP:20140918T214606
LAST-MODIFIED:20131002T104309
CREATED:20130827T120000
SEQUENCE:5
CATEGORIES:DRO
DTSTART:20130920T163000Z
DTEND:20130920T183000Z
UID:9f136f46-0f2b-11e3-9e19-00259064d38a8272013_11:16:17_AM@gsb.columbia.edu
SUMMARY;ENCODING=QUOTED-PRINTABLE:DRO PhD Seminar: Yonatan Gur (DRO) & Ningyuan Chen (IEOR)
LOCATION:Uris 1st Floor - 142, 3022 Broadway, New York, NY 10027
URL:http://groups.gsb.columbia.edu/rsvp.aspx?id=173984
DESCRIPTION:\nSpeaker: Yonatan Gur. . Title: Non-stationary Stochastic Optimization. . Coauthors: O. Besbes, Y. Gur, and A. Zeevi. . Abstract: We consider a non-stationary variant of a sequential stochastic optimization problem, where the underlying cost functions may change along the horizon. We propose a measure, termed variation budget, that controls the extent of said change, and study how restrictions on this budget impact achievable performance. We identify sharp conditions under which it is possible to achieve long-run-average optimality and more refined performance measures such as rate optimality that fully characterize the complexity of such problems. In doing so, we also establish a strong connection between two rather disparate strands of literature: adversarial online convex optimization; and the more traditional stochastic approximation paradigm (couched in a non-stationary setting). This connection is the key to deriving well performing policies in the latter, by leveraging structure of optimal policies in the former. Finally, tight bounds on the minimax regret allow us to quantify the "price of non-stationarity," which mathematically captures the added complexity embedded in a temporally changing environment versus a stationary one.. . . . . . Speaker: Ningyuan Chen. . Title: Directed random graphs with given degree distributions. . Abstract: In the talk we present a model for generating directed random graphs with given degree distributions. We first draw degree sequences and make the sum of in- and out-degress equal, without modifying the degree distribution significantly. Then we construct the graph by the directed configuration model. The resulted graph is guaranteed to be simple by repeating the algorithm, if the degree distributions have finite variance, or by removing the self-loops and multiple edges, if the degree distributions have infinite variance. Joint work with Mariana Olvera.. . \nEvent Organizer: Karin Eriksson (kbe2105@columbia.edu)\n---\nRSVP: http://groups.gsb.columbia.edu/rsvp.aspx?id=173984\n---\n[EVENT_TYPE:Research]\n[GROUP_TYPE:Academic Offices]\n
BEGIN:VALARM
TRIGGER:-PT15M
ACTION:DISPLAY
DESCRIPTION:Reminder
END:VALARM
END:VEVENT
END:VCALENDAR