BEGIN:VCALENDAR VERSION:2.0 PRODID:researchseminars.org CALSCALE:GREGORIAN X-WR-CALNAME:researchseminars.org BEGIN:VEVENT SUMMARY:Jannis Kurtz (University of Siegen) DTSTART:20200714T180000Z DTEND:20200714T183000Z DTSTAMP:20260423T021753Z UID:DOTs/21 DESCRIPTION:Title: Di screte Optimization Methods for Group Model Selection in Compressed Sensin g\nby Jannis Kurtz (University of Siegen) as part of Discrete Optimiza tion Talks\n\n\nAbstract\nWe study the problem of signal recovery for grou p models. More precisely for a given set of groups\, each containing a sma ll subset of indices\, and for given linear sketches of the true signal ve ctor which is known to be group-sparse in the sense that its support is co ntained in the union of a small number of these groups\, we study algorith ms which successfully recover the true signal just by the knowledge of its linear sketches. We consider two versions of the classical Iterative Hard Thresholding algorithm (IHT). The classical version iteratively calculate s the exact projection of a vector onto the group model\, while the approx imate version (AM-IHT) uses a head- and a tail-approximation iteratively. We apply both variants to group models and analyse the two cases where the sensing matrix is a Gaussian matrix and a model expander matrix.\n\nTo so lve the exact projection problem on the group model\, which is known to be equivalent to the maximum weight coverage problem\, we use discrete optim ization methods based on dynamic programming and Benders' Decomposition. T he head- and tail-approximations are derived by a classical greedy-method and LP-rounding\, respectively.\n LOCATION:https://researchseminars.org/talk/DOTs/21/ END:VEVENT END:VCALENDAR