08+09S 09W 10S,W 11S,W 12S,W 13S,W 14S,W 15S,W 16S,W 17S,W 18S,W 19S,W 20S,W 21S,W 22S,W 23S,W

Summer Term 2023, Doctoral School Events
2023-03-31 Doctoral School Seminar (10:30–13:00, TU, Kopernikusg. 24)
Daodao Yang (TU, advisor R. Tichy): Extreme values of Dirichlet polynomials with multiplicative coefficients [show abstract]
Lea Enzi (TU, advisor S. Thonhauser): Numerical computation of risk functionals [show abstract]
Bianca Dornelas (TU, advisor M. Kerber): Sparse higher order Čech complexes exist [show abstract]
2022-04-21 Doctoral School Seminar (DATE HAS CHANGED) (15:00–17:30, KFU, Heinrichstr. 30)
Rodolfo Assereto (KFU, advisor K. Bredies): A Multiscale-Proximal Algorithm for Total-Variation Denoising [show abstract]
New Students (TU): P. Tabatabai, M. Henry
Enis Chenchene (KFU, advisor K. Bredies): A hybrid proximal generalized conditional gradient method and application to total variation parameter learning [show abstract]
2023-05-26 Doctoral School Seminar (10:45–13:00, TU, Kopernikusg. 24)
Dominik Schmid (TU, advisor M. Kang): Catching a robber on a random hypergraph

Abstract: The game of Cops and Robber is a two-player game played on a graph, in which a group of cops attempt to catch a robber moving along the edges of the graph. The cop number of a graph is the minimum number of cops required to win the game. In this talk, we consider the game of Cops and Robber on a hypergraph, where the players move along hyperedges instead of edges. We conjecture that the cop number of a k-uniform hypergraph on n vertices is O(√(n/k)) and prove this conjecture up to logarithmic factors asymptotically for random hypergraphs

Maximilian Ofner (TU, advisor S. Hörmann): Using factor models to reconstruct partially observed functional data

Abstract: We study linear reconstruction of functions from noisy measurements on a proper subset of the functions' domain. To this end, we propose a new estimation approach which is based on approximate factor models of increasing rank. Under mild regularity conditions, we derive uniform convergence rates of our estimator using a triple asymptotic. Finally, the methodology is illustrated with a simulation study and a set of real temperature data.

Dominik Lang (KFU, advisor G. Leobacher): Optimal Staking of Proof-of-Stake Crypto Assets

Abstract: We introduce a novel problem of optimizing the staking of Proof-of-Stake (PoS) cryptocurrencies by extending classical portfolio optimization methods. In PoS investors can voluntarily lock their coins into a smart contract, restricting their liquidity for a specific period, in exchange for earning rewards within the PoS cryptocurrency network. We propose a Markov Decision Problem framework for the investor and discuss difficulties in finding analytic solutions. Finally, we employ methods from reinforcement learning for numerical optimization.