EF1 – Extracting dynamical Laws from Complex Data

Project

EF1-16

Quiver Representations in Big Data and Machine Learning

Project Heads

Alexander Schmitt

Project Members

Jorge Arturo Esquivel Araya, Jan Marten Sevenster

Project Duration

01.04.2022 − 31.03.2025

Located at

FU Berlin

Description

Quiver representations arise, e.g., as persistence modules in big data or as sets of weights of a neural network. We will combine algorithmic and algebro geometric methods in order to classify quiver representations arising from big data and to analyze the geometry of the moduli space of a neural network.

Related Publications

A. Schmitt: An elementary discussion of the representation and geometric
invariant theory of equioriented quivers of type D with an application to
quiver bundles. Linear Multilinear Algebra.
https://www.tandfonline.com/doi/full/10.1080/03081087.2021.1983512

 

M.A. Armenta Armenta, A. Schmitt: The Hilbert-Mumford criterion for representations of network quivers. In Proceedings of the International Conference on Applied Sciences and Engineering (ICASE 2023). Atlantis Highlights in Engineering.
https://www.atlantis-press.com/proceedings/icase-23/125996825

 

Related Pictures

ICASE 2023

Alexander Schmitt at the conference ICASE, Ulaanbataar, 2023