BEGIN:VCALENDAR
VERSION:2.0
METHOD:PUBLISH
CALSCALE:GREGORIAN
PRODID:-//WordPress - MECv7.15.0//EN
X-ORIGINAL-URL:https://dcn.nat.fau.eu/
X-WR-CALNAME:
X-WR-CALDESC:FAU DCN-AvH. Chair for Dynamics, Control, Machine Learning and Numerics -Alexander von Humboldt Professorship
REFRESH-INTERVAL;VALUE=DURATION:PT1H
X-PUBLISHED-TTL:PT1H
X-MS-OLK-FORCEINSPECTOROPEN:TRUE
BEGIN:VEVENT
CLASS:PUBLIC
UID:MEC-fc1671e392a4646fdc594a1e58ac635c@dcn.nat.fau.eu
DTSTART:20220926T085000Z
DTEND:20220926T103000Z
DTSTAMP:20220926T102800Z
CREATED:20220926
LAST-MODIFIED:20240924
PRIORITY:5
SEQUENCE:1
TRANSP:OPAQUE
SUMMARY:Frontiers in Mathematics: Control and Machine Learning by E. Zuazua
DESCRIPTION:On Monday September 26, 2022 our Head Prof. Enrique Zuazua will talk on “Control and Machine Learning” as invited speaker at the Frontiers in Mathematics, 2022 International Conference on the Cooperation and Integration of Industry, Education, Research and Application from September 26th to 29th, 2022 at School of Mathematics, Jilin University (Changchun, China).\nAsbtract. In this lecture we shall present some recent results on the interplay between control and Machine Learning, and more precisely, Supervised Learning and Universal Approximation. We adopt the perspective of the simultaneous or ensemble control of systems of Residual Neural Networks (ResNets). Roughly, each item to be classified corresponds to a different initial datum for the Cauchy problem of the ResNets, leading to an ensemble of solutions to be driven to the corresponding targets, associated to the labels, by means of the same control.\nWe present a genuinely nonlinear and constructive method, allowing to show that such an ambitious goal can be achieved, estimating the complexity of the control strategies. This property is rarely fulfilled by the classical dynamical systems in Mechanics and the very nonlinear nature of the activation function governing the ResNet dynamics plays a determinant role. It allows deforming half of the phase space while the other half remains invariant, a property that classical models in mechanics do not fulfill. The turnpike property is also analyzed in this context, showing that a suitable choice of the cost functional used to train the ResNet leads to more stable and robust dynamics.\nThis lecture is inspired in joint work, among others, with Borjan Geshkovski (MIT), Carlos Esteve (Cambridge), Domenec Ruiz-Balet (IC, London) and Dario Pighin (Sherpa.ai)\nWHERE\nTalk: Control & Machine Learning by E. Zuazua (10:50H -Berlin time)\nZoom meeting ID: 862 062 0549 | PIN code: 2022\nThe conference in this year focuses on the following four topics: Numerical Methods and Analysis, Algebra and Applications, Dynamical Systems and Differential Equations with Applications, Control, Optimization and Data Science.\n
URL:https://dcn.nat.fau.eu/events/frontiers-in-mathematics-control-and-machine-learning-by-e-zuazua/
CATEGORIES:EZuazua,Seminar/Talk,Workshop
END:VEVENT
END:VCALENDAR