SearcharxivSearch

arXiv subjects

Elias Sink

Publications and source records attributed to Elias Sink.

2 recordsLinked to original sources

Line configurations and K3 surfaces

We study the realization spaces of $10_3$ line configurations. Answering a question posed by Sturmfels in 1991, we use elliptic surface techniques to show that realizations over $\mathbb{Q}$ are dense in those over $\mathbb{R}$ for all $10_3$ configurations. We find that for exactly four of the ten configurations, the realization space admits a compactification by a K3 surface. We show that these have Picard number 20 and compute their discriminants. Finally, we use geometric invariant theory to give an elegant interpretation of these K3 surfaces as moduli spaces.

math.AG

Noncommutative resolution of $SU_C(2)$

We study the derived category of the moduli space $SU_C(2)$ of rank $2$ vector bundles on a smooth projective curve $C$ of genus $g\ge 2$ with trivial determinant. This generalizes the recent work by Tevelev and Torres on the case with fixed odd determinant. Since $SU_C(2)$ is singular, we work with its resolution of singularities, specifically with the noncommutative resolution constructed by Pădurariu and Špenko--Van den Bergh (in the more general setting of symmetric stacks). We show that this noncommutative resolution admits a semiorthogonal decomposition into derived categories of symmetric powers $Sym^{2k}C$ for $2k\le g-1$. In the case of even genus, each block appears four times. This is also true in the case of odd genus, except that the top symmetric power $Sym^{g-1}C$ appears twice. In the case of even genus, the noncommutative resolution is strongly crepant in the sense of Kuznetsov and categorifies the intersection cohomology of $SU_C(2)$. Since all of its components are "geometric," our semiorthogonal decomposition provides evidence for the expectation, which dates back to the work of Newstead and Tyurin, that $SU_C(2)$ is a rational variety. Finally, we study mutations of semiorthogonal decompositions on the Hecke correspondence, answering a question of Pădurariu and Toda.

math.AG