- This event has passed.

# The geometry of coherent sheaves: From derived categories to Higgs bundles

## April 24, 2023 at 15:00 – 17:30 CEST

GAUS-Workshop: “Recent developments in GIT”

14:00-15:00: **Victoria Hoskins** (Nijmegen, speaking remotely): An introduction to geometric invariant theory

Abstract: The aim of this survey talk is to give an introduction to geometric invariant theory in order to prepare the audience for the subsequent talks as requested by the organisers. I will start by explaining how group actions often appear in moduli problems and we will see how constructing algebra-geometric quotients is related to 19th century invariant theory. I will explain why the theory is simplest for non-reductive group actions and, in this case, I will explain how Mumford constructs quotients (of certain open ‘semistable’ subsets) using geometric invariant theory, as well as giving combinatorial and numerical criteria for semistability. If there is time, I will briefly mention some recent developments to extend GIT to certain non-reductive group actions.

15:20-16:20: **Joshua Jackson **(Sheffield): Advances in Non-reductive GIT and applications

*
*Abstract: Following from the previous talk on reductive GIT, I will survey recent developments in extending this theory to non-reductive groups, with a particular focus on applications to moduli theory. Time permitting, I will then indicate how non-reductive GIT can be used in the study of sheaves, Higgs bundles, hypersurfaces, and singular curves.

16:40-17:40: **Dario Weissmann **(Essen): A stacky approach to identify the semi-stable locus of vector bundles

Abstract: I report on recent joint work with Xucheng Zhang focusing on our Theorem A for vector bundles in characteristic 0: The semi-stable locus in the stack of bundles over a smooth projective curve is the maximal open locus admitting a schematic good moduli space. This gives an intrinsic motivation for semi-stability of vector bundles. Historically, semi-stability appeared in the quest for a moduli space of bundles and the classical construction of this moduli space uses a non-canonical GIT-construction. Theorem A also provides us with natural examples of good moduli spaces which are only algebraic spaces and not schemes.

* *