New approaches to finite generation of cohomology rings (Q1984357): Difference between revisions

For any associative algebra $B$ and a $B$-bimodule $M$, one has a morphism of graded commutative algebras $\varphi_M: HH^*(B)\to \mathrm{Ext}_B^*(M,M)$ from Hochschild cohomology of $B$ to the ordinary $B$-cohomology of the pair $(M,M)$. The bilateral ideal $ker(\varphi_M)$ is called \textit{the support variety of $M$}. Support varieties have been a subject of active research since the seminal work of \textit{D. Quillen} on the spectrum of equivariant cohomology rings [Ann. Math. (2) 94, 549--572, 573--602 (1971; Zbl 0247.57013)]. One of the important questions that the theory of support varieties aims to answer is to decide whether $HH^*(A)$ is finitely generated, or better yet, whether $HH^*(A,M)$ is noetherian over $HH^*(A)$ for every finitely generated $A$-bimodule $M$. Throughout this review we will call the latter condition as \textit{the Noetherian hypothesis} that the authors abbreviate as \textbf{(fg)} in the article. The cohomology ring $HH^*(A)$ is finitely generated when the Noetherian hypothesis holds, but the hypothesis is known to be false in general [\textit{S. Zhu}, J. Algebra 180, No. 1, 187--205 (1996; Zbl 0845.16039)]. The theory of support varieties was originally developed for group rings. Group rings form a particular and a large class of Hopf algebras. Moreover, for any Hopf algebra $A$ there is a canonical isomorphism between $HH^*(A,M)$ and $H^*(A,M^{\mathrm{adj}}) = \mathrm{Ext}_A^*(k,M^{\mathrm{adj}})$ the ordinary cohomology of $A$ with coefficients in $M$ viewed as a $A$-module via the adjoint action. The latter contains $H^*(A,k)$ as a subalgebra when we set $M=A$. In the article under review, the authors transport the Noetherian hypothesis question to a analogous question over $H^*(A,k)$, and investigate under which conditions $H^*(A,M)$ is a Noetherian over $H^*(A,k)$. They prove the Hopf variant of the Noetherian hypothesis, which they abbreviate as \textbf{(hfg)}, when \begin{itemize} \item[1.] $A$ has the integral property (Section 3), \item[2.] $A$ has permanent cocycles (Section 4), and \item[3.] $A$ has good a reduction modulo $p$, for some prime $p$ (Section 5). \end{itemize}