On the classification of quantum symmetries

TL;DR

This paper develops a classification framework for quantum symmetries via Hopf algebra coactions, providing new results on the structure of actions on finite-dimensional algebras and revealing limitations of universal coacting Hopf algebras.

Contribution

It introduces a classification approach based on cosupports, classifies quantum symmetries of specific algebras, and demonstrates new bounds on regrading finite-dimensional matrix algebras.

Findings

Classifies quantum symmetries of two points and dual numbers.

Shows the non-existence of a universal coacting Manin Hopf algebra on a straight line.

Establishes that for n ≥ 14, certain matrix algebras admit nontrivial coactions with unique properties.

Abstract

We show that, in order to classify Hopf algebra (co)actions on a given finite dimensional algebra up to equivalence, one should start with the classification of the possible cosupports (i.e. the sets of linear operators by which $H^*$ is acting) of Hopf algebra coactions and then consider dual Hopf algebra actions. As an application, we classify quantum symmetries of the set of two points and the algebra of dual numbers. In addition, we show that the straight line does not admit an (ungraded) universal coacting Manin Hopf algebra. Moreover, we prove that for $n\geqslant 14$ the full matrix algebra $M_n(\mathbb k)$ admits a nontrivial Hopf algebra coaction such that all Hopf algebra actions with the same restriction on the cosupport are trivial, i.e. the cosupport may reduce under the dualization and a finite dimensional algebra may have less equivalence classes of actions than coactions. Simultaneously, for any $n\geqslant 14$, we define an elementary grading on $M_n(\mathbb k)$ by an infinite group that cannot be regraded by any finite group. (The previously known lower bound was $n\geqslant 349$.)