Optimization seminar

Modal title

Modal content

Datum / Zeit

Referent:in

Titel

Ort

26. September 2016
16:00-17:00

Dr. David Adjiashvili
Institute for Operations Research of ETH Zurich

Details

Titel	Improved Approximation for Weighted Tree Augmentation with Bounded Costs (discussion seminar)
Referent:in, Affiliation	Dr. David Adjiashvili, Institute for Operations Research of ETH Zurich
Datum, Zeit	26. September 2016, 16:00-17:00
Ort	HG G 19.1

Improved Approximation for Weighted Tree Augmentation with Bounded Costs (discussion seminar)

17. Oktober 2016
16:00-17:00

Dr. Stefan Weltge
Institute for Operations Research of ETH Zurich

Details

Titel	Extended formulations: Constructions (discussion seminar)
Referent:in, Affiliation	Dr. Stefan Weltge, Institute for Operations Research of ETH Zurich
Datum, Zeit	17. Oktober 2016, 16:00-17:00
Ort	HG G 19.1

Extended formulations: Constructions (discussion seminar)

24. Oktober 2016
16:00-17:00

Silvia Di Gregorio
Università degli Studi di Padova, Padova, Italia

Details

Titel	Structure of extreme functions with discrete domain
Referent:in, Affiliation	Silvia Di Gregorio, Università degli Studi di Padova, Padova, Italia
Datum, Zeit	24. Oktober 2016, 16:00-17:00
Ort	HG G 19.1

Structure of extreme functions with discrete domain

31. Oktober 2016
16:00-17:00

Dr. Stephen Chestnut
Institute for Operations Research of ETH Zurich

Details

Titel	Concentration (discussion seminar)
Referent:in, Affiliation	Dr. Stephen Chestnut, Institute for Operations Research of ETH Zurich
Datum, Zeit	31. Oktober 2016, 16:00-17:00
Ort	HG F 26.3
Abstract	In probability theory, a concentration inequality bounds how a random variable X deviates from its mean. These inequalities are extremely important for the study of randomized algorithms. When X is a sum of independent random variables, we can apply our old favorites like Chebyshev's and Chernoff's inequalities to bound the deviation of the sum from its mean. But, what if X is more complicated than a simple sum? Can we still prove that it concentrates? Very often, the answer is yes. We'll start the talk with some variants of the famous Johnson-Lindenstrauss Lemma and I'll show you how one concentration inequality plays a role in a fast randomized algorithm for linear least squares. Next I'll introduce some more general methods for proving concentration inequalities, with a focus on the Efron-Stein Inequality and the "Entropy Method". We'll see how these work by applying them to prove concentration for a monotone submodular function evaluated on a random set.

Concentration (discussion seminar)read_more

Hinweise: wenn Sie möchten, können Sie den iCal/ics-Kalender abonnieren.

Archiv: FS 22 FS 20 HS 19 FS 19 HS 18 FS 18 HS 17 FS 17 HS 16 FS 16 HS 15 FS 15 HS 14 FS 14 HS 13 FS 13 HS 12 FS 12 HS 11 FS 11 HS 10 FS 10 HS 09