BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//TYPO3/NONSGML Calendar Base (cal) V1.11.0//EN
METHOD:PUBLISH
BEGIN:VEVENT
UID:www.mysite.com_1_5803
DTSTAMP:20200309T095233
DTSTART:20200630T140000Z
DTEND:20200630T163000Z
CATEGORIES:Kolloquium
SUMMARY:Deep learning concepts for inverse problems - Regularization by architecture -*Abgesagt
DESCRIPTION:Deep learning concepts are presently intruding almost all fields of science and engineering. Suchconcepts produce astonishing experimental results and seem to bypass easily well established andwell researched classical approaches in particular for applications in data analysis and image processing. However\, this is not the case for inverse problems where applying deep neural networks asa generic toolbox fails.\r\n\r\nThe classical approach to inverse problems starts with an analytical description F : X -> Yof theforward operator in some function spaces X; Y . The field of inverse problems addresses the task ofreconstructing an unknown x from noisy data y ~ F(x) with the further complication that the inverse of F or any type of generalized inverse is unbounded.\r\n\r\nThis inherent and unavoidable instability\, which cannot be remedied by preconditioning or any othertype of data preprocessing\, is reflected by the failure of naively transferring deep learning conceptsfrom image processing directly to inverse problems in tomography\, non-destructive testing or monitoring physical-technical processes in general. In this talk we will discuss specific deep learning concepts for inverse problems\, which allow an interpretation in terms of the classical analytical regularization theory. These resuts so far only apply to comparatively small network designs. In addition we demonstrate the potential for large scale problems in the field of magnetic particle imaging
LOCATION:Campus Grifflenberg, Gebäude G, Ebene 15, Raum 20, Gaußstr. 20
ORGANIZER;CN="Fakultät für Mathematik und Naturwissenschaften, Fachgruppe Mathematik und Informatik":
END:VEVENT
END:VCALENDAR