On the modular isomorphism problem for groups with center of index at most $p^3$

Extracted source text · arXiv 2311.06666

\documentclass[10pt]{amsart}
\usepackage{amsmath,amssymb,verbatim}
\usepackage[utf8]{inputenc}
%\usepackage[latin1]{inputenc}
%\usepackage{latexcad,eDepic,epic}
\usepackage[english]{babel}
\usepackage{enumerate}
\usepackage{pdflscape}
\usepackage{amsthm}
%\usepackage{stmaryrd}
\usepackage{mathrsfs}
%\usepackage[dvips]{graphicx}
\usepackage{color}
\usepackage{multicol}
\usepackage[normalem]{ulem}
\usepackage{cancel}
\usepackage{tikz}
\usetikzlibrary{matrix}
\usepackage[all]{xy}
\usepackage{mathtools}
\usepackage{arydshln}
\usepackage[shortlabels]{enumitem}
 \usepackage[english]{babel}
% \usepackage[notref,notcite]{showkeys}
\usepackage{hyperref}
\hypersetup{
    colorlinks = true,
    linkbordercolor = {white},
    allcolors = blue
} 
\usepackage[capitalize, noabbrev]{cleveref}
 
\makeatletter
\@namedef{subjclassname@1991}{\textup{2020} Mathematics Subject Classification}
\makeatother

\topmargin=-.15in %Negative number is measured from the default position
\textheight=9.0in %Top-to-bottom length of printed area
\topskip=\baselineskip \flushbottom %\headsep=0in
%\headheight=0in
\textwidth=6.5in %Width of printed area %\textwidth=5.6in %for better 
%screen display
\evensidemargin=0in %0in from default position.
\oddsidemargin=0in




%\usepackage{amsmath,amssymb}
%\renewcommand{\theenumi}{\alph{enumi}}
\renewcommand{\labelenumi}{(\theenumi)}

\newtheorem{theorem}{Theorem}[section]
\newtheorem{lemma}[theorem]{Lemma}
\newtheorem{fact}[theorem]{Fact}
\newtheorem{proposition}[theorem]{Proposition}
\newtheorem{example}[theorem]{Example}
\newtheorem{definition}[theorem]{Definition}
\newtheorem{corollary}[theorem]{Corollary}
\newtheorem{remark}[theorem]{Remark}
\newtheorem{examples}[theorem]{Examples}
\newtheorem{remarks}[theorem]{Remarks}
\newtheorem{notation}[theorem]{Notation}
\newtheorem{problem}[theorem]{Problem}
\newtheorem{conjecture}{Conjecture}
\newtheorem{algorithm}{Algorithm}

\newtheorem*{theorem*}{Theorem}

\newtheorem{maintheorem}{Theorem}
\newtheorem{maincorollary}[maintheorem]{Corollary}

\renewcommand\themaintheorem{\Alph{maintheorem}}

%\newcommand{\qed}{\enspace\vrule height6pt width4pt depth2pt}

%\def\thedefinition{\null}
%\def\thetheorem{\arabic{theorem}}
\makeatletter
\@addtoreset{equation}{section}
\makeatother
\def\thetheorem{\thesection.\arabic{theorem}}
\def\theequation{\thesection.\arabic{equation}}

%\newenvironment{proof}{\par\noindent{\bf Proof.}}{$\qed$\par\bigskip}



\newcommand\doverline[1]{\overline{\overline{#1}}}
\newcommand\toverline[1]{\overline{\doverline{#1}}}


%comandos nuevos
\newcommand{\ChAngel}[1]{{\bf \textcolor{blue}{#1}}} %This is the reason of the Ch(ange) commands. 
\newcommand{\ChDiego}[1]{{\bf \textcolor{purple}{#1}}}
\newcommand{\ChMima}[1]{{\bf \textcolor{orange}{#1}}}
\newcommand{\ChSofia}[1]{{\bf \textcolor{orange}{#1}}} %Choose your colour. 

\newcommand{\nsp}{\upharpoonleft \hspace{-4.5pt}\downharpoonleft}
\newcommand{\M}{{\mathrm{D}}}
\newcommand{\bo}{{\mathcal{O}}}
\newcommand{\bi}{{\mathcal{I}}}
\newcommand{\D}{{\mathcal{D}}}
\newcommand{\q}{{\mathcal{Q}}}
\newcommand{\bb}{{\overline b}}

\newcommand{\Gal}{{\rm Gal}}
\newcommand{\vcd}{{\rm vcd}}
\newcommand{\cd}{{\rm cd}}
\newcommand{\soc}{{\rm soc}}
\newcommand{\mcd}{\mbox{\rm gcd}}
\newcommand{\Cen}{{\rm C}}
\newcommand{\core}{\mbox{\rm Core}}
\newcommand{\Aut}{\mbox{\rm Aut}}
\newcommand{\Inn}{\mbox{\rm Inn}}
\newcommand{\Ker}{\mbox{\rm Ker }}
\newcommand{\Imagen}{\mbox{\rm Im }}
\newcommand{\Gr}{\mbox{\rm Gr}}
\newcommand{\StGr}{\mbox{\rm StGr}}
\newcommand{\CP}{\mbox{\rm CP}}
\newcommand{\Pic}{\mbox{\rm Pic}}
\newcommand{\Out}{\mbox{\rm Out}}
\newcommand{\Orb}{\mbox{\rm Orb}}
%\newcommand{{\bf H}om}{\mbox{\rm Hom}}
\newcommand{\Id}{\mbox{\rm Id}}
\newcommand{\AntiHom}{\mbox{\rm AntiHom}}
\newcommand{\End}{\mbox{\rm End}}
%\newcommand{\mod}{\mbox{\rm mod}}
\newcommand{\gr}{\mbox{\rm gr}}
\newcommand{\sma}{R \# P_G}
\newcommand{\Stab}{\mbox{\rm Stab}}
\newcommand{\inv}{\textup{inv}}
\newcommand{\INV}{\text{INV}}
\newcommand{\N}{{\mathbb N}}
\newcommand{\Z}{{\mathbb Z}}
\newcommand{\Q}{{\mathbb Q}}
\newcommand{\R}{{\mathbb R}}
\newcommand{\C}{{\mathbb C}}
\newcommand{\F}{{\mathbb F}}
\newcommand{\HQ}{{\mathbb H}}
\newcommand{\K}{{\mathbb K}}
\newcommand{\flecha}{\rightarrow}
%\newcommand{\U}[1]{#1^�}
\newcommand{\G}{\mathcal{G}}
\newcommand{\B}{\mathcal{B}}
\newcommand{\Gm}{G_{m,n_1,n_2}^{s_1,s_2}}
\newcommand{\res}{\mbox{res}}
\newcommand{\GL}{{\rm GL}}
\newcommand{\SL}{{\rm SL}}
\newcommand{\SGL}{{\rm SGL}}
\newcommand{\SO}{{\rm SO}}
\newcommand{\nr}{{\rm nr}}
\newcommand{\SG}{{\rm S}\Gamma}
\newcommand{\PSL}{{\rm PSL}}
\newcommand{\PE}{{\rm PE}}
\newcommand{\tr}{{\rm tr}}
\newcommand{\Soc}{{\rm Soc}}
\newcommand{\ind}{{\rm ind}}

\newcommand{\matriz}[1]{\begin{array} #1 \end{array}}
\newcommand{\pmatriz}[1]{\left(\begin{array} #1 \end{array}\right)}
\newcommand{\GEN}[1]{\left\langle #1 \right\rangle}
\newcommand{\rk}{{\rm rk}}
\newcommand{\WH}[1]{\widehat{#1}}
\newcommand{\gorro}[1]{\widehat{#1}}
\newcommand{\MC}[1]{\mathcal{#1}}
\newcommand{\ov}[1]{\overline{#1}}
\newcommand{\aug}[1]{\mathrm{I}(#1)}
\newcommand{\augNor}[2]{\mathrm{I}(#1;#2)}
\newcommand{\Aug}[2]{\mathrm{J}_k^{#1}(#2)}
\newcommand{\mJ}[1]{\mathcal{J}_{#1}}
\newcommand{\mK}[1]{\mathcal{K}_{#1}}

\newcommand{\quat}[2]{\left( \frac{#1}{#2} \right)}
\newcommand{\I}{{\bf i}}
\renewcommand{\j}{{\bf j}}
\renewcommand{\k}{{\bf k}}
%\newcommand{\e}[1]{\left\lfloor #1 \right\rfloor}
%\newcommand{\E}[1]{\left\lceil #1 \right\rceil}

\newcommand{\ZZ}{\mathrm{Z}}
\newcommand{\U}{\mathcal{U}}
%\newcommand{\V}{\mathcal{V}}
\newcommand{\T}{\mathcal{T}}
\newcommand{\W}{\mathcal{W}}
\newcommand{\CC}{\mathcal{C}}

\newcommand{\e}[1]{\left\lfloor #1 \right\rfloor}
\newcommand{\E}[1]{\left\lceil #1 \right\rceil}


\newcommand{\V}{\mathrm{V}}

\newcommand{\Ci}{$\blacktriangleleft$}
\newcommand{\Cf}{$\blacktriangleright$}

\newcommand{\va}{\ov{a}}
\newcommand{\vb}{\ov{b}}
\newcommand{\ord}{\text{o}}
\newcommand{\lex}{\text{lex}}
\newcommand{\VSd}{V(S_0)}
\newcommand{\VSdd}{V}

\newcommand{\Ese}[2]{\mathcal{S}\left(#1\mid #2\right)}
\newcommand{\Te}[2]{\mathcal{T}\left(#1\mid #2\right)}


\newenvironment{proofof}{\medskip\par\noindent \textit{Proof of }}{\qed\par\bigskip}

\newcommand{\pa}[2]{\varepsilon_{#2}(#1)}

\newcommand{\diag}{\operatorname{diag}}
\DeclareMathOperator{\Cl}{Cl}
\DeclareMathOperator{\Hom}{Hom}
\DeclareMathOperator{\Irr}{Irr}
\DeclareMathOperator{\IBr}{IBr}
\DeclareMathOperator{\IRR}{IRR}
\DeclareMathOperator{\VPA}{VPA}
\newcommand{\qand}{\quad \text{and} \quad}
\newcommand{\Supp}{{\rm Supp}}
\DeclareMathOperator{\Ch}{Ch}
\DeclareMathOperator{\PCh}{PCh}
\DeclareMathOperator{\DCh}{DCh}
\DeclareMathOperator{\Cyc}{C}
\DeclareMathOperator{\Cocyc}{Cocyc}
\DeclareMathOperator{\PAP}{PAP}
\DeclareMathOperator{\Exp}{Exp}
\DeclareMathOperator{\OO}{O}
\DeclareMathOperator{\dG}{d}


\newcommand{\indA}[1]{\ind_{(H,#1)}^{H\times N}(1)}
\newcommand{\inda}[2]{\ind_{(H_{#2},#1)}^{H_{#2}\times N_{#2}}(1)}
\newcommand{\INDA}[3]{\ind_{(#2,#1)}^{#2\times #3}(1)}

%\title[On the Modular Isomorphism Problem for  two-generated cyclic-by-abelian groups]{On the Modular Isomorphism Problem for  two-generated cyclic-by-abelian groups}

\title[On the modular isomorphism problem for groups with center of index at most $p^3$]{On the modular isomorphism problem  \\ for groups with center of index at most $p^3$}
 
 \author{Sofia   Brenner}
\address{Department  of  Mathematics,   TU  Darmstadt,   Germany.}
\email{brenner@mathematik.tu-darmstadt.de}
 
 \author{Diego García-Lucas}
 \address{Departamento de Matem\'aticas, Universidad de Murcia, Spain}
 \email{diego.garcial@um.es}

\thanks{The research of the first author has received funding from the European Research Council (ERC) under
	the European Unions Horizon 2020 research and innovation programme (EngageS: grant agreement No.~820148). The research of the second author was partially supported by Grant PID2020-113206GB-I00 funded by MCIN/AEI/10.13039/501100011033  and by Grant Fundación Séneca 22004/PI/22.}


\keywords{Finite $p$-groups, modular group algebra, small group algebra,  modular isomorphism problem.}

\subjclass{20C05, 16S34}

\date{\today}



\begin{document}
	
	\maketitle
	
	\begin{abstract}
	Let $p$ be an odd prime number. We show that the modular isomorphism problem has a positive answer for finite $p$-groups whose center has index $p^3$, which is a strong contrast to the analogous situation for $p = 2$.
	\end{abstract}

%As $G'$ is elementary abelian, if $G$ has nilpotency class $2$, then it satisfies the MIP by \cite{San89}. Thus we assume that $G$ has nilpotency class $3$. In particular $G/\ZZ(G)$ is non-abelian. Denote $d(G)$ the minimal number of generators of $G$, i.e., $p^{d(G)}=|G/\Phi(G)|$. If $G$ is $2$-generated, then $G$ satisfies the MIP by a   commentary without proof at the end of \cite{Bag99}, proved in \cite[Proposition 1.4]{MM20}. Thus we may assume that $d(G)\geq 3$. On the other hand for these kind of groups $d(G)$ cannot be too big (I'd say $d(G)\leq 5$, but I have not a proof yet.) Moreover, since $\gamma_2(G)^p \gamma_4(G)=1$, if also $[G:C_G(G')]=p$ and $C_G(G')$, then $G$ satisfies the MIP by \cite[Theorem 3.3]{MS22}. Thus we assume otherwise, i.e., either $[G:C_G(G')]>p$ or $C_G(G')$ is not abelian. Let $\M_i(G)=G\cap (1+I(\F_p G)^i$, the $i$-th term of   the Jennings series of $G$. If  $\M_4(G)=1$, then $G$ satisfies the MIP by \cite{Her07}, so we assume otherwise. We can also assume that $G$ has no abelian direct factors by \cite{GL2022}. Summarizing:
% 
% \begin{lemma}
% 	We can assume that\begin{itemize} 
% 		\item $\M_4(G)\neq 1$.
% 		\item $G/\ZZ(G)$ has order $p^3$ and it is non-abelian.
% 		\item $\gamma_2(G)^p\gamma_4(G)=1$.
% 		\item Either $C_G(G')$ is not maximal or $C_G(G')$ is not abelian. 
% 		\item $d(G)\geq 3$.
% 		\item $G$ has no abelian direct factors. 
% 	\end{itemize}
% \end{lemma}
%
%	I also conjecture that, in our situation,  $G/\ZZ(G)\cong (C_3\times C_3)\rtimes C_3$, the SmallGroup in GAP with Id [27,3], and that the other non-abelian group cannot occur. 
%	
%	
%	At this point I started to worry, because until here everything was kind of trivial. However, from now on it is not going to be so nice. I calculated the groups of order $3^7$ satisfying all our hypotheses, and there are exactly $7$ of them; moreover only $2$ of them cannot be distinguished by the group-theoretical invariants implemented in the GAP package  ModIsomExt. Some of these invariants can be calculated with the help of a computer, but in a theoretical argument they might be impossible to use. The remaining groups are [ 2187, 559 ] and [ 2187, 583 ].	It is known that the groups of order $3^7$  satisfy the  MIP, though this is done by computational means and nobody knows exactly where is the difference between the group algebras. So a good first step is understanding this difference for this concrete pair of groups, with the hope that it can be translated to other cases. 
%	
%	
%Anyways	I think the number of groups in our conditions sharing all the known invariants will grow if we increase the order of the group, so this  class of groups is not MIP-trivial. 
%	
%
%
%One of the techniques used to obtain the results I mentioned is to study the so-called \emph{small group algebra} $S=\F_p G/I(G)I(G')$. It is known that the isomorphism type of $S$ depends on the isomorphism type of $\F_p G$ alone.   If $V(S)$ is the group of units of $S$ of augmentation $1$, then in our situation 
%$$V(S)=G\rtimes A $$
%for some abelian $p$-group $A$ by \cite[Proposition 3.2]{MM20}. I don't know if this will be useful, but we can study the proof of \cite[Proposition 1.4]{MM20}, and see if something can be recovered. One first option we have is to follow this path, with the hope that the isomorphism type of $G$ can be somehow read from $V(S)$. 
%
%
%Another path would be to classify up to isomorphism the groups in our hypotheses, and once the groups are understood, try to use the known invariants (see the list \cite[Theorem 2.1]{MM20}, or the ones that can be obtained from \cite[Lemmas 3.2 and 3.6]{GL2022} -- read just the statements, I don't think the proof these last lemmas will be useful). Maybe you can check also Leo's survey \cite{Mar22}.  
%
%
%
%\section{Update 29-06}
%
% \begin{fact}
% 	$G/\ZZ(G)=(C_p\times C_p)\rtimes C_p$.
% \end{fact}
%\begin{proof}
%	Otherwise $G/\ZZ(G)\cong C_{p^2}\rtimes C_p$. Let $a,b\in G$ such that $a\ZZ(G)$ has order $p^2$, $b\ZZ(G)$ has order $p$, and $[a\ZZ(G), b\ZZ(G)]=a^p \ZZ(G)$. Then $G'\ZZ(G)=\GEN{a^p,\ZZ(G)}$, so $C_G(G')=C_G(a^p)=\GEN{a, \ZZ(G)}$, which has index $p$ in $G$, contradicting our hypothesis. 
%\end{proof}
%
%In the following we fix $a,b\in G$ such that $a\ZZ(G)$ and $b\ZZ(G)$ generate $G/\ZZ(G)$, and write $c=[a,b]$. 
%\begin{fact} $|G'/\gamma_3(G)|=p$ and 
%	$|G'|\leq p^3$.
%\end{fact}
%\begin{proof} 
%	As $C_G(G')=G'\ZZ(G)$ has index $p^2$ in $G$, $|G'\ZZ(G)/\ZZ(G)|$ has order $p$  and the bilinear map 
%	$$[,]:G/C_G(G') \times G'\ZZ(G)/\ZZ(G) \to \gamma_3(G)$$
%	is surjective, we derive that $|\gamma_3(G)|\leq p^2$. On the other hand, $G/G'\ZZ(G)\times G/G'\ZZ(G)\to G'/\gamma_3(G)$ is also surjective, and 
%	$  [a^xb^y G' \ZZ(G), a^z b^tG'\ZZ(G)] =c^{xt-yz} \gamma_3(G)$, we derive that $|G'/\gamma_3(G)|=p$, and the lemma follows. 
%\end{proof}
%
%Let $z_1,\dots,z_l \in \ZZ(G)$ such that $\{a,b,z_1,\dots, z_l\}$ form a Burnside basis of $G$.  Moroever $a$, $b$ and the $z_i$'s can be chosen such that $G/G'=\GEN{aG'}\times \GEN{bG'}\times \GEN{z_1G'}\times \dots \times \GEN{z_lG'}\cong C_{p^{n}}\times C_{p^m}\times C_{p^{n_1}}\times \dots \times  C_{p^{n_l}}$. Then  
%$$G/\gamma_3(G)=\GEN{a\gamma_3(G),b \gamma_3(G)} \times \GEN{z_1\gamma_3(G)}\times \dots \times \GEN{z_l \gamma_3(G)} ,$$
%as $ [a,b] \notin \ZZ(G)             $. Still $\GEN{z_i\gamma_3(G)}\cong C_{p^{n_i}}$. Moreover we can reorder the $z_i$'s so that $n_i\geq n_{i+1}$, and the same with $a$ and $b$ to assume that $n\geq m$.
%
%
%\begin{fact}\label{fact:boundGens}
%	$d(G)\leq 4$, and  if $|\gamma_3(G)|=p$  then $d(G)\leq 3$. 
%\end{fact}
%\begin{proof}
%	Suppose by contradiction that $d(G)>4$. Then $l\geq 3$. Suppose  that $|\gamma_3(G)|=p^2$. Then $\gamma_3(G)=\GEN{c_1}\times \GEN{c_2}$ is elementary abelian, so we can regard it as a vector space over $\F_p$.   %Write $z_i^{p^{n_i}}=c_1^{\delta_{1i}} c_2^{\delta_{2i}}$, with $0\leq \delta_i \leq p-1$.
%	 As $z_1^{p^{n_1}}$, $z_2^{p^{n_2}}$ and $z_3^{p^{n_3}}$ are linearly dependent, we can write $z_3^{p^n_3}= z_1^{\varepsilon_1 p^{n_1}} z_2^{\varepsilon_2 p^{n_2}}$. Then substitute $z_3$ by $z_3 z_1^{-\varepsilon_1 p^{n_1-n_3}} z_2^{-\varepsilon_2 p^{n_2-n_3}}$. Then for the new $z_3$ we have that $\GEN{z_3}\cap \gamma_3(G)=1$ and $z_3$ is central, so $\GEN{z_3}$ is a direct factor of $G$. Contradiction. A completely analogous argument shows the remaining statement. 
%\end{proof}


%\newpage

\section{Introduction} 

Let $p$ be a prime number and let $F$ be the field with $p$ elements. In this paper, we study the modular isomorphism problem, which can be stated as follows: \medskip

\emph{Consider finite $p$-groups $G$ and $H$ such that the group algebras $FG$ and $FH$ are isomorphic. Is it true that $G$ and $H$ are isomorphic?}
\medskip


%\ChSofia{should we comment on the choice of the field?}\ChDiego{[I don't think so: our result is restricted to the prime field, as is the original statement of the problem.]}

In the past 50 years, this problem has seen extensive research. There are many classes of finite $p$-groups for which it is known to have a positive answer, such as the class of abelian  $p$-groups \cite{Deskins1956}, the class of $p$-groups  of nilpotency class $2$ with elementary abelian derived subgroup \cite{San89}, or the class of metacyclic groups \cite{BaginskiMetacyclic, San96}.
  The modular isomorphism problem has a positive answer for groups with a center of index $p^2$. This follows from the earlier-mentioned result by Sandling \cite{San89}. Moreover, the generalization of the problem for this class of groups to arbitrary fields of characteristic $p$ has a positive solution, as shown by    \cite{Drensky}. The modular isomorphism problem is also known to have positive answer for groups of order dividing $p^5$ for any prime \cite{Passman1965p4, SalimSandlingp5}, for groups of order dividing $2^8$, for groups of order dividing $3^7$ and, except for a few groups, for groups of order dividing $5^6$ \cite{Eick08, MM20}. %\ChSofia{I like that.}
 Some recent results are for example \cite{BK19,MS22}, and a complete overview can be found in the recent survey~\cite{Mar22}. 
\medskip

Despite all positive results, it was shown in \cite{GarciaMargolisdelRio} that the general answer to this problem is negative. The smallest counterexample consists of a pair of groups of order $2^9$ with centers of index $8$.  For $p > 2$, the modular isomorphism problem is still open.
\medskip

 For a $p$-group $G$, we write $\ZZ(G)$ for its center, $\Phi(G)$ for its Frattini subgroup and $(\gamma_i(G))_{i \geq 1}$ for its lower central series. Motivated by the counterexample of \cite{GarciaMargolisdelRio} for $p = 2$, we now study finite $p$-groups with a center of index $p^3$ for $p > 2$. In contrast to the situation for $p = 2$, it turns out that the modular isomorphism problem has a positive answer for this class of groups. Our main result is the following: 
 
\begin{maintheorem}\label{theo:maintheorema}
	Let $p>2$ be a prime integer, let $F$ be the field with $p$ elements, and let $G$ and $H$ be  finite $p$-groups. Suppose that $|G:\ZZ(G)|=p^3$. If $FG\cong FH$, then $G\cong  H$. 
\end{maintheorem} 

Note that the class of $p$-groups $G$ with $|G:\ZZ(G)| = p^3$ is not contained in the union of the classes of $p$-groups for which the modular isomorphism problem is known to have a positive answer. Corresponding examples are discussed in \Cref{sec:indexp3}. 
The key ingredient for the proof of \Cref{theo:maintheorema} is the following result, which might also be of independent interest: 


                                                                                                                  
\begin{maintheorem}\label{theo:maintheoremb}                                                                                
Let $p>2$ be a prime integer, let $F$ be the field with $p$ elements, and consider finite $p$-groups~$G$ and $H$ with $F G\cong F H$. Let $d \in \Z_{\geq 0}$ with $|G/\Phi(G)\ZZ(G)|=p^d$. If 
\begin{enumerate}
		\item   $G^p\cap \gamma_2(G)\subseteq \gamma_2(G)^p \gamma_3(G)$ and 
		\item  $|\gamma_2(G)/\gamma_2(G)^p\gamma_3(G)|=p^{{d\choose 2}} ,$
	\end{enumerate}
	then $G/\gamma_2(G)^p \gamma_4(G)\cong  H/\gamma_2(H)^p \gamma_4(H)$.
\end{maintheorem}

This statement is a generalization of \cite[Theorem~1.2]{MM20} when $p>2$, as for $d=2$, it is equivalent to the following implication: if $ G/\ZZ(G )$ is $2$-generated, then the isomorphism class of $G/\gamma_2(G)^p\gamma_4(G)$ is determined by $FG$.
\medskip

This paper is organized in the following way: In \Cref{sec:preliminaries}, we introduce the notation used in this paper. In \Cref{sec:smallgroupalgebra}, we derive results on the so-called small group algebra, which will lead to the proof of~\Cref{theo:maintheoremb}. Finally, in \Cref{sec:indexp3}, we prove \Cref{theo:maintheorema}. 


 



\section{Preliminaries and notation}\label{sec:preliminaries}
Let $p$ be an odd prime number and let $G$ be a finite $p$-group. We use the standard group-theoretic notation. In particular, $Z(G)$ and $\Phi(G)$ denote the center and the Frattini subgroup of $G$, respectively. A \emph{Burnside basis} is a generating system of $G$ of minimal size. 
For $a,b \in G$, let $[a,b] = a^{-1} b^{-1} a b$. Moreover, we set $\gamma_1(G) = G$ and $\gamma_{i+1}(G) = [\gamma_i(G), G]$ for $i\geq 1$. 

\begin{lemma}\label{lemma:computation}
	Let $G$ be a finite $p$-group of nilpotency class at most $3$. Then for each $x,y\in G$, 
	$$[x^p,y] = [x,y]^p [[x,y] ,x]^{{p\choose 2}} \qand [x,y^p]=[x,y]^p [[x,y],y]^{{p\choose 2}}.$$
\end{lemma}
\begin{proof}
	Using several times the identity $$[xz,y]=[x,y]^z [z,y]=[x,y][[x,y],z] [z,y], \qquad ( x,y,z\in G) $$   we derive that	\begin{align*}
	[x^p,y]&=  [x,y]^{x^{p-1}}\cdot [x,y]^{x^{p-2}}\cdots [x,y] \\
	&=[x,y] [[x,y],x]^{p-1}\cdot [x,y]^{x^{p-2}}[[x,y],x]^{p-2}\cdots [x,y] \\
%	&= [x,y]^p [[x,y],x]^{\sum_{i=1}^{p-1} i} \\
	&= [x,y]^p [[x,y],x]^{ {p \choose 2}}.
	\end{align*}   
	A similar argument shows the second identity. 
\end{proof}
As an immediate consequence, we obtain:
\begin{lemma}\label{lemma:AgemoCentral}
	Let $G$ be a finite $p$-group with $p>2$. If $\gamma_2(G)^p\gamma_4(G)=1$, then $G^p\subseteq \ZZ(G)$.
\end{lemma}


 
 Let $F$ be a field of characteristic $p$. In this paper we only consider unitary, finite-dimensional $F$-algebras.
 An \emph{augmented} $F$-algebra is an $F$-algebra $A$ endowed with an $F$-algebra homomorphism  $A\to F$, called the \emph{augmentation map}. The group algebra $FG$ is an augmented $F$-algebra with augmentation map $\sum_{g\in G} a_gg\mapsto  \sum_{g\in G} a_g$. The kernel $I(A)$ of the augmentation map is called the \emph{augmentation ideal} of~$A$. For $A=FG$, we abbreviate $I(FG)$ by $I(G)$.

 Let $A$ be an augmented local $F$-algebra, as it is the case for $A=FG$. Then we denote $V(A)= 1+I(A)$, the subgroup of the group of units of $A$ formed by the elements of augmentation $1$.  Every quotient of $A$ by some ideal $J$ of $A$ is also an augmented algebra with an augmentation map induced by the augmentation of~$A$. We can (and will)  make the identifications $I(A/J)=I(A)/J$ and $V(A/J)=1+I(A/J)=V(A)/(1+J)$.
We define the set $K(A)$ to be the $F$-subspace of $A$ generated by the  elements of the form $ab-ba$ with  $a,b\in A$.
	
	
%	\ChDiego{[What do you think of the previous paragraph? If we keep it (improvements are welcome), we should delete from here]}\ChSofia{I like it}
	
%Let $F$ be a field of characteristic $p > 2$. For $a,b \in FG$, we define the commutator $[a,b] = ab - ba$ and set $K(FG)$ to be the $F$-subspace generated by the commutators of the elements in $FG$. Usually, it will become clear from the context whether the group-theoretic or the ring-theoretic notion of commutators is meant. We consider the \emph{augmentation map} 
%\[FG \to F,\, \sum_{g \in G} a_g g \mapsto \sum_{g \in G} a_g,\]
%which is a ring homomorphism. Its kernel, the \emph{augmentation ideal} of $G$, is denoted by $I(G)$. Let $V(FG)$ denote the \emph{group of normalized units} of $FG$, that is, $V(FG) = 1 + I(G)$. \ChDiego{ Maybe we should set the  notation for augmented algebras in general, then then just use the $V(\cdot)$ operator? For example: Given an augmented local algebra $A$, we denote $I(A)$ its augmentation ideal and $V(A)\coloneqq 1+ I(A)$ its group of units of augmentation $1$. When $A=FG$, we abbreviate $I(FG)=I(G)$. Every quotient of $A$ by some ideal $J$ of $A$ is also an augmented algebra, and we can (and will)  make the identifications $I(A/J)=I(A)/J$ and $V(A/J)=1+I(A/J)=V(A)/(1+J) $. }
%\ChSofia{we can do that, but we should make clear that for the group algebra, we always want the particular augmentation map given above (I guess that's a priori not clear)}
%
%\ChDiego{[Until here]}

Note that $I(G) I(\gamma_2(G))$ is a $2$-sided ideal of $FG$. A main tool in this paper will be the investigation of the \emph{small group algebra} $S = FG/I(G) I(\gamma_2(G))$. This object has been extensively studied by R.~Sandling and M.~A.~M.~Salim (see \cite{San89, SalimSandling1995, SalimTesis}).  It was used in~\cite{SalimSandlingp5} to positively answer the modular isomorphism problem for groups of order dividing $p^5$. Moreover, the analogous construction for   group rings over the integers was used in  A.~Whitcomb's thesis \cite{Whitcomb} to positively answer the isomorphism problem  for integral group rings for metabelian groups.   Further results about the small group algebra in a wider sense can be found in   \cite{HertweckSoriano07}.  
%We write $V(S)=1+ I(S)=1+ I(G)/I(G)I(\gamma_2(S))$ for its group of units of augmentation $1$. 
By a bar, we denote both the natural projection $\overline\cdot :FG\to S$, which is an $F$-algebra homomorphism, and its restriction $\overline \cdot :V(FG)\to V(S)$, which can be viewed as a group homomorphism. A subgroup of $V(FG)$ that is an $F$-basis for $FG$ is called a \emph{group basis} of $FG$. 
We say a property $P$ is \emph{determined} by the group algebra if for any pair of $p$-groups $G$ and $H$ such that $FG\cong FH$, the group $G$ satisfies $P$ if and only if $H$ satisfies $P$.
%\ChSofia{List of possible things to define/recall: $I(G)$, $\gamma_i$, augmentation, (Frattini subgroup), $V(FG)$. Maybe we should introduce a short notation for the nilpotency class. \\}
%\ChDiego{[Is this lemma in Huppert's book?]\ChSofia{ I didn't find it. But it is the kind of exercise result you can accidentally restate without being accused of copying, I guess. }}\ChDiego{Ok.}

We have that $(1+I(G)I(\gamma_2(G)))\cap G=\Phi(\gamma_2(G))$   (see the introduction of \cite{SalimSandling1995}), so $\overline{G} \cong G/\Phi(\gamma_2(G))$. Note that this holds for any group basis of $FG$. However, it has been known for a long time that the information encoded by the small group algebra alone is not enough to recover the isomorphism type of $G/\Phi(\gamma_2(G))$ (see \cite[Section 4]{Bag99}), even for groups of nilpotency class $3$ (see \cite[Example 3.11]{MS22} and the subsequent discussion). 


We frequently use the following result due to Sandling:  

\begin{lemma}[{\cite{San89}, see also \cite[Proposition~3.2]{MM20}}]\label{lemma:definitiona}
	%	\ChSofia{it looks quite ugly with the $\pi(G)$, nicer ways to write it down?} \ChDiego{[In can't think a better notation. I propose substitute $\pi'$ and $\pi''$ for $\pi_1$ and $\pi_2$, but this does not solve the ugliness of $\pi(G)$ in this proposition. ]}  \ChSofia{I think with the $\bar{G}$, it now looks bearable} \ChDiego{[I'm okay with the notation. And definitively looks better.]}
	Let $G$ be a finite $p$-group. 
	Let $g_1, \ldots, g_m \in \overline{G}$ be elements whose images in $\overline{G}/\gamma_2(\overline{G})$ form an independent generating set as an abelian group. Let $A$ be the subgroup of $V(S)$ generated by the elements $1 + (g_1-1)^{k_1} \cdots (g_m-1)^{k_m}$ with $k_1, \ldots, k_m \in \Z_{\geq 0}$ and $k_1 + \dots + k_m \geq 2$. Then $V(S) = \overline{G}A$. For $a = 1 + (g_1-1)^{k_1} \cdots (g_m-1)^{k_m} \in A$ and $g \in \overline{G}$, we have \[[g,a] = [g, g_1, g_1, \dots, g_1, g_2, \dots, g_2, \dots, g_m, \dots, g_m],\] where each $g_i$ appears exactly $k_i$ times. If additionally $\gamma_4(G) = 1$ holds, then $A$ is abelian and $V(S) = \overline{G} \rtimes A$. 
\end{lemma}

Furthermore, \cite[Theorem~1.1]{SalimSandling1995} yields that 
\begin{equation}\label{eq:gammaivs}
\gamma_i(V(S))=\gamma_i(\overline{G})
\end{equation} 
for each $i\geq 2$. 

\section{Results on the small group algebra}\label{sec:smallgroupalgebra}
%\ChSofia{potentially move background info on the small group algebra to the previous section?}
%\ChSofia{prerequisites for the results?} \ChDiego{[Ok. And then the title of \Cref{sec:smallgroupalgebra} becomes Proof ot Theorem B?]}\ChSofia{can also be something like Results on the small group algebra - whatever you like}
Throughout, let $F$ denote the field with $p$ elements for a prime number $p> 2$ and let $G$ be a finite $p$-group. In this section, we derive results on the small group algebra that will later be used to solve the modular isomorphism problem for groups with a center of index~$p^3$. 

We use the notation introduced in the preceding section. In particular, let $\overline{\cdot} \colon FG\to S$ be the natural projection onto the small group algebra $S$ and let $V(S) = 1 + I(G)/I(G) I(\gamma_2(G))$ denote the group of units of augmentation $1$ of $S$. Furthermore, let $A$ be a subgroup of $V(S)$ as described in Lemma~\ref{lemma:definitiona}. 

%By a result of Sandling \cite{San89}, as explained in \cite[Proposition~2.4]{MM20}, there exists a subgroup $A$ of $S$ such that $S= \pi(G) A$. 

\begin{remark}
For every normal subgroup $N$ of $G$ contained in $\gamma_2(G)$, it follows easily  from the identity $xy-1=x-1+y-1+(x-1)(y-1)$  that \[I(N)FG+ I(\gamma_2(G))I(G)= N-1+I(\gamma_2(G))I(G).\] 
\end{remark}

In particular, $\gamma_i(\overline G)-1 = \overline{\gamma_i(G)-1}$ is an ideal in $S$.
%\begin{lemma}
%For every $i \geq 2$, the set $\gamma_i(\bar G)-1$ is an ideal in $S$.
%\end{lemma}
%
%\begin{proof}
%Fix $i \geq 2$. In this proof, we write $[a,b] = ab - ba$ for the commutator of elements in $a,b \in FG$, and $[x,y]_G = x^{-1} y^{-1} xy$ for the commutator of group elements $x, y\in G$. 
%For $x,y \in \gamma_i(G)$, we have $(x-1) + (y-1) = (xy-1)-(x-1)(y-1) \in \Gamma$. In particular, $\gamma_i(G)-1$, and hence $\gamma_i(\bar{G})-1$, is additively closed. Now let $x\in \gamma_i(G)$ and $y \in I(G)$. Then $y(x-1) \in I(G) I(\gamma_2(G))$ follows. In particular, we have $\overline{y(x-1)} \in \bar{G}-1$. On the other hand, note that $(x-1)y = y (x-1) + [y,x-1]$. Write $y = \sum_{g \in G} a_g (g-1)$ with $a_g \in F$. Then
%\begin{alignat*}{2} 
%[y,x-1] &= \sum_{g \in G} a_g [g-1,x-1] \\
%&= \sum_{g \in G} a_g [g,x] \\
%&= \sum_{g \in G} a_g (1 + g-1 + x-1 + (g-1)(x-1)) ([x,g]_G-1) \\
%&\equiv \sum_{g \in G} (a_g [x,g]_G-1 )\pmod{I(G) I(\gamma_2(G))}.
%\end{alignat*}
%Since $\gamma_i(\bar{G})-1$ is closed under addition, we obtain $\overline{[y,x-1]} \in \gamma_i(\bar{\Gamma})-1$, which concludes the proof. 
%\end{proof}
%
%\ChSofia{maybe at least rename $\pi'$ to $\pi$ (which we don't have anymore) and $\pi''$ to $\pi'$, but maybe a different name would be more suggestive. I am also wondering whether we need $\pi'$ explicitly, or whether we could do both in one projection¸}
%\ChSofia{I renamed all quotients of the small group algebra to something with $S$ and the unit groups to something containing $V$, but I am very open to alternatives} 
%\ChDiego{[The advantages of having first $\pi'$ and $\pi''$ is that the first is an homomorphism of rings, while the second one is only an homomorphism of groups. I want to avoid the notation things+prime, so I'm okay with any change you make in that sense.] }
%\ChSofia{I just realized that the projection onto $S'$ should be $\pi'$, so forget about my comment above. I haven't come up with nicer ideas so far. For the time being, I renamed $V_{S''}$ to a simple $V$, but we can think of alternatives}
%\ChSofia{further think about notation here}
In the following, we set $\Gamma = \gamma_4(G) -1 \subseteq FG$ and consider the quotient algebra  $S_0 = FG/(I(G) I(\gamma_2(G)) + \Gamma)  \cong S/\overline{\Gamma}$ together with the corresponding projection $\pi_0\colon FG \to S_0$.
Then $\pi_0(G)= G/\gamma_2(G)^p \gamma_4(G)$, and the same holds for every group basis of~$FG$. %We set $\VSd=1+\pi'(I(G))$. 
Furthermore, we obtain $\VSd = \pi_0(G) \rtimes A$ (see \Cref{lemma:definitiona}). %for the subgroup $A$ defined in \Cref{lemma:definitiona}. 
More generally, the following holds:
%Then the previous yield $S'=\pi'(G) A$ for the subgroup $A$ defined in \Cref{lemma:definitiona}. Moreover, \cite[Proposition 2.4\,(iv)]{MM20} guarantees that $S'=\pi'(G)\rtimes A$. 
%\begin{lemma}\label{lemma:orders}
%	Let $G$ and $H$ finite $p$-groups such that $FG\cong FH$. Then $|G/\Phi(\gamma_2(G))|=|H/\Phi(\gamma_2(H))|$ and $|G/\gamma_2(G)^p\gamma_4(G)|=|H/\gamma_2(H)^p\gamma_4(H)|$.
%\end{lemma}
%\begin{proof}
%	This is well-known. The first equality follows as both $|\gamma_2(G)|$ and $|\gamma_2(G) /\Phi(\gamma_2(G))|$ are determined by the group algebra. The second one follows follows from the previous and the identity $\gamma_4(\pi(G))=\gamma_4(S)=\gamma_4(\pi(H))$.
%\end{proof}

 
  
 \begin{lemma}\cite[Proposition 3.1]{MS22}\label{lemma:complement}
 	Let $H$ be any group basis of $FG$. Then $\VSd=\pi_0(H)\rtimes A$.
 \end{lemma}
% \begin{proof}
% 	As $S'=\pi'(G)\rtimes A$ and $\gamma_2(S')=\gamma_2(\pi'(G))$, we have that $$S/\Phi(S)=G\Phi(S)/\Phi(S)\times A\Phi(S)/\Phi(S)\cong  G/\Phi(G)\times A/\Phi(A)$$
% 	As $A\subseteq 1+I(FG)^2$, we have	 a short exact sequence
% 	$$1\to A \Phi(S)/\Phi(S)\to S/\Phi(S)\to S/1+I(FG)^2\to 1 ,$$
% 	where (the image of) a Burnside basis of $G$ (or $H$) is still a Burnside basis for the elementary abelian group $S/1+I(FG)^2 =1+I(FG)/1+I(FG)^2$.  Hence $S/\Phi(S)= H\Phi(S)/\Phi(S)\times A\Phi(S)/\Phi(S)$.  This implies that $S=\pi(H)A$, and as $|S|=|\pi(G)||A|=|\pi(H)||A|$ by \Cref{lemma:orders}, and $H$ is normal in $S$, it follows that $S=\pi(H)\rtimes A$.
% \end{proof}


%We use the theory of Jennings (especially Jennings bases and Jennings tuples)  as explained in \cite[Section 2.3]{MS22}. 
%\begin{lemma}\label{lemma:JTN}
%	Let $N$ be a normal subgroup of a finite $p$-group $G$. Then there is a Jennings tuple $\mathscr T$ of $G$ such that $N\cap \mathscr T$ is a Jennings  tuple for $N$. 
%\end{lemma}
%\begin{proof}
%	We prove the first assertion by induction on $|N|$.  If $|N|=1$, there is nothing to prove. Now suppose that the result holds for normal subgroups of order $p^n$ for some $n\geq 0$, and assume that $N$ has order $p^{n+1}$. Since $G$ is a $p$-group, the center of $G$ intersects $N$ non-trivially, so we can choose a subgroup $L\subseteq N\cap \ZZ(G)$ of order~$p$. Then, by the induction hypothesis, we can choose a Jennings tuple $\bar {\mathscr T}$ of $G/L$ such that $\bar  {\mathscr T}\cap (N/L)$ is a Jennings tuple for~$N/L$. Let $\mathscr T$ be a set of preimages of the elements of $\bar {\mathscr T}$ in $G$. Clearly, the preimages of elements in $N/L$ are in~$N$. For some $i$ we have that  $L \subseteq D_i(G)$  but $L \not \subseteq D_{i+1}(G)$, and for some $j$, that $   L\subseteq \M_j (N)$ but $L\not\subseteq \M_{j+1}(N)$. Observe that $\mathscr T$ is almost a Jennings basis of $G$ except it does not contain preimages of a basis of $\M_i(G)/\M_{i+1}(G)$, only of a maximal linear subspace which is a direct complement of $L$. Similarly, $\mathscr T\cap N$ is almost a Jennings basis of $N$ except it does not contain preimages of a basis of $\M_j(N)/\M_{j+1}(N)$, only of a maximal linear subspace which is a direct complement of $L$. Hence it suffices to take the Jennings tuple  $\mathscr T\cup \{l\}$, where $l$ is a generator of $L$. 
%	
%	  
%\end{proof}
%
% 
% 
% \ChDiego{[I think lemma 2.3 is useless for us here, and Lemma 2.4's proof is incorrect.  ]}
%\begin{lemma}
%	  $\pi(1+\ZZ(FG)\cap I(G))\subseteq  \ZZ(\pi(G))  (A\cap \ZZ(S))$.
%\end{lemma}
%\begin{proof}
%	Let $\mathscr T $ be a Jennings tuple  of $G$ such that  $\mathscr T_0=\mathscr T\cap \ZZ(G)$ is a Jennings tuple for $\ZZ(G)$, which is possible by \Cref{lemma:JTN}. An element $x$ of augmentation $1$ in $\ZZ(FG)=F\ZZ(G)+ \ZZ(FG)\cap [FG,FG]$  is of the form $x=1+z+ \alpha$, where $z\in  I(\ZZ(G))$ and $\alpha $ is linear combination of elements of the form $\sum_{g\in C} g$, where~$C$ is a non-central conjugacy class of $G$. For each such class $C$, we have $\sum_{g\in C} g= g\sum_{d\in D} d$ for some fixed $g\in  C$ and some subset $D\subseteq G'$ such that $|D|=|C|$. The latter is a non-trivial power of $p$. Hence
%	\begin{align*}                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           
%		\sum_{g\in C}g+I(G)I(G')&= g\sum_{d\in D} (d-1)+I(G)I(G') \\                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
%		&= g (\prod_{d\in D} d-1)+I(G)I(G') \\                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     
%		&= \prod_{d\in D} d-1 +(g-1)(\prod_{d\in D} d-1) +  I(G)I(G') \\                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           
%		&=  \prod_{d\in D} d-1 +I(G)I(G') ,                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        
%	\end{align*}                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                
%where we use the identity $xy-1=x-1+y-1+(x-1)(y-1)$. The same identity guarantees that $\alpha=c-1$ modulo $I(G) I(G')$, for some $c\in G'$. On the other hand, by   \cite[Proposition 2.6]{San84Ab}, $1+z$ is of the form $h  \eta  $, where  $\eta$ is a product of elements of the form $\prod_{g\in \mathscr T_0}(g-1)^{k_g}$ with $\sum_{g\in \mathscr T_0} k_g\geq 2$, $k_g\in \Z_{\geq 0}$, and $h\in \ZZ(G)$.  Thus $ x+I(G)I(G')= h\eta  + c-1+I(G)I(G')=   h \eta d  +I(G)I(G')$. Observe that $\pi(\eta)$ lies in $A$ and is clearly central, while $\pi(hc)$ is in the image of $G$, and also has to be central, so the result follows.                                                                                                                                  
%\end{proof}
%As a straightforward consequence, we get:
%\begin{lemma}
%	$\pi'(1+\ZZ(FG)\cap I(G))\subseteq \ZZ(\pi'(G))\times (A\cap \ZZ(S'))$.
%\end{lemma}

Finally, we consider the natural projection from $\VSd$ onto $\VSdd= \VSd/ (A\cap \ZZ(\VSd))$. Composing it with the restriction of~$\pi_0$ to $V(FG)$ yields a group homomorphism 
\begin{equation}\label{eq:defpi}
\pi \colon V(FG)\to \VSdd.
\end{equation} Hence $\VSdd =\pi(G)\rtimes E$ follows, where $E=A/ (A\cap \ZZ(\VSd))$ is elementary abelian due to $A^p\subseteq \ZZ(\VSd)$ (see \Cref{lemma:definitiona}). Therefore, by \Cref{lemma:complement}, we have 
\begin{equation}\label{eq:decompv}
	V = \pi(H) \rtimes E
\end{equation}
  for any group basis $H$ of $FG$. 
%  \ChSofia{I think the decomposition of $V$ is a bit hidden here, we use it in the proof of Theorem 4.4. Maybe phrase as a lemma/remark, so that we can put a reference?}\ChDiego{Maybe just a numbered equation?} \ChSofia{also fine for me -- I think it makes people happy to click on a hyperlink if they forgot the definition of symbols :D}
%This means that $\pi''(1+\ZZ(FG)\cap I(G))\subseteq \ZZ(\pi''(G))$, and by the same reason $\pi''(1+\ZZ(FG)\cap I(G))\subseteq \ZZ(\pi''(H))$. 

%\ChSofia{maybe motivate the next lemma - otherwise it might frighten people}\ChDiego{[Ok, but how? On the other hand, I have nothing against scaring people a bit.]} \ChSofia{Okayyyy... I made an attempt with the following sentence.}
In the next lemma, we derive a presentation for groups satisfying the hypothesis of Theorem~\ref{theo:maintheoremb}. 

 \begin{lemma}\label{lemma:presentation}  
	Write $|G/\ZZ(G)\Phi(G)|=p^d$, $|\gamma_2(G)^p \gamma_3(G)/\gamma_2(G)^p\gamma_4(G)|=p^m$ and $|G/\Phi(G)|=p^k$. If \begin{enumerate}[(a)]
		\item  $G^p\cap \gamma_2(G)\subseteq \gamma_2(G)^p \gamma_3(G)$ and 
		\item $|\gamma_2(G)/\gamma_2(G)^p\gamma_3(G)|=p^{{d\choose 2}} $,
	\end{enumerate}  then there exist positive integers $n_i$  and  $0\leq \alpha_{is}, \beta_{ijls}\leq p-1$ for   $i,j,l\in \{1,\dots, k\}$ and $s\in \{1,\dots, m\}$ such that %\ChDiego{[I modified the notation for $\mathcal R$: now it is uglier but more precise.]} 
$$G/\gamma_2(G)^p \gamma_4(G)\cong \GEN{g_1,\dots, g_k ,c_1,\dots, c_m \mid \mathcal R_{ (n_i,\alpha_{ij},\beta_{ijls})} ( g_1,\dots, g_k,c_1,\dots, c_m)   },$$   where $\mathcal R_{( n_i,\alpha_{ij},\beta_{ijls})} ( g_1,\dots, g_k,c_1,\dots, c_m) $ is the set formed by the following relations: 
\begin{enumerate}
	\item \label{1} $g_i^{p^{n_i}}=\prod_{1\leq j \leq m}c_j^{\alpha_{ij}}$   for each   $1\leq i  \leq k$.
	\item \label{2}$ [[g_i,g_j], g_l]=\prod_{1\leq s\leq m} c_s^{\beta_{ijls}}$ for $1\leq i,j,l\leq d$.
	\item\label{3} $[g_i,g_j]^p=1$ for  $1\leq i ,j\leq d$.
	\item\label{4} $g_i$ is central for $d<i\leq k$. 
	%\ChSofia{maybe we should write (4) and (5) as proper relations, this is a bit sloppy} \ChDiego{[If you prefer so, go ahead. My impresion is that this way was friendlier to the reader.]} \ChSofia{Maybe we leave it that way. Can we write a full sentence? ``The nilpotency class of $G$ is at most 3. Would feel less sloppy to me for some irrational reason.} \ChDiego{[Done as you propose, except it is not ``of $G$''; maybe we can write ``of the group'' instead?]} \ChSofia{Done}
	\item\label{5} The nilpotency class of the group is at most $3$.
\end{enumerate} 
Moreover, we can choose the elements $g_{d+1},\dots ,g_{k} \in G/\gamma_2(G)^p \gamma_4(G)$ such that they have central preimages in $G$.
\end{lemma}
%	\ChSofia{I did not come up with something instead of $\tilde{G}$ so far}\ChDiego{[Ángel suggested --I forgot-- widetildes and overlines over single letters, so they can be more easily distinguished.]} \ChSofia{I changed that -- what do you think?}\ChDiego{[I think it is fine.]}
\begin{proof}
Set $\widetilde{G}=G/\gamma_2(G)^p \gamma_4(G)$ and, for $g \in G$, write $\widetilde{g} = g \gamma_2(G)^p \gamma_4(G) \in \widetilde{G}$ (similarly for subsets of~$G$). Observe that $|\gamma_3(\widetilde{G})|=p^m$ and $|\widetilde{G}/\Phi(\widetilde{G})|=p^k$. Write $ |\widetilde{G}/\ZZ(\widetilde{G}) \Phi(\widetilde{G})|=p^t$. As the image of $Z(G) \Phi(G)$ in $\widetilde{G}$ has index at most~$p^d$, we obtain   
$p^t\leq p^d$.  Moreover, $\gamma_2(\widetilde G)$ is elementary abelian and the hypotheses imply that $ \widetilde {G}^p \cap\gamma_2(\widetilde G)\subseteq \gamma_3(\widetilde G)  $ and $|\gamma_2(\widetilde G)/\gamma_3(\widetilde G)|=p^{d\choose 2}$.  Since ${\widetilde G}^p\subseteq \ZZ(\widetilde{G})$ by \cref{lemma:AgemoCentral}, we have a   well-defined map $\mathfrak c \colon \widetilde{G}/\ZZ(\widetilde{G}) \Phi(\widetilde{G}) \times \widetilde{G}/\ZZ(\widetilde{G}) \Phi(\widetilde{G}) \to  \gamma_2(\widetilde G)/\gamma_3(\widetilde G) $ induced by the group commutator map, which is bilinear and anti-symmetric. If $\{x_1,\dots, x_t\}$ is a basis of $\widetilde{G}/\ZZ(\widetilde{G}) \Phi(\widetilde{G})$, then $\mathfrak c(x_i,x_j)$ with $1\leq i <j \leq t$ generates $\gamma_2(\widetilde G)/\gamma_3(\widetilde G)$ as an $F$-vector space. Thus $p^{{d\choose 2}}=|\gamma_2(\widetilde G)/\gamma_3(\widetilde G)|   \leq p^{{t\choose 2}}$. This yields $t=d$. 
 


Let $\{x_1,\dots, x_d\}$ be a set of preimages of a basis of $  G/\ZZ(  G)\Phi(  G)$ in $G$, and let $\{x_{d+1},\dots x_k$\} be a set of central preimages in $G$ of a basis of $\ZZ(G)\Phi(G)/\Phi(G)$. Thus $\{x_1,\dots, x_k\}$ is a Burnside basis of $G$, and its image is a Burnside basis of $\widetilde G$. By abuse of notation, we regard  $\{x_1,\dots, x_k\}$ as a subset of $\widetilde G$. Let $\{z_1,\dots , z_m\}$ be a basis of $\gamma_3(\widetilde G)$. Write $p^{n_i}$ for the order of $x_i \gamma_2(\widetilde{G})$ in $\widetilde{G}/\gamma_2(\widetilde G)$. As $\widetilde {G}^p\cap \gamma_2(\widetilde G) \subseteq \gamma_3(\widetilde G)$, we have that $x_i^{p^{n_i}}=\prod_{1\leq j \leq m} z_j^{\alpha_{ij}}$ for uniquely determined $\alpha_{ij} \in \{0, \dots, p-1\}$. Moreover, for $1\leq i,j,l\leq d$, we can write $ [[x_i,x_j], x_l]=\prod_{1\leq s\leq m} z_s^{\beta_{ijls}}$ for uniquely determined $\beta_{ijls} \in \{0, \dots, p-1\}$. As $\{z_1, \ldots, z_m\}$ is a basis of $\gamma_3(\widetilde G)$, the matrix $(\beta_{ijls})_{ (i,j,l), s}$,   with rows indexed by triples $(i,j,l)$ and columns indexed by~$s$, has rank $m$. 
	
	Let \[H=\GEN{g_1,\dots, g_k,c_1,\dots ,c_m \mid \mathcal R_{ (n_i,\alpha_{ij},\beta_{ijls})} ( g_1,\dots, g_k,c_1,\dots, c_m) }.\] 	 The assignment $g_i \mapsto x_i$ ($i = 1, \ldots, k$) and $c_j \mapsto z_j$ ($j = 1, \ldots, m$) defines a surjective group homomorphism $H\to \widetilde G$. To complete the proof it suffices to show that it is an isomorphism, or equivalently that $|H|\leq  |\widetilde G|$. 
	 
	
	 Observe that since the matrix $(\beta_{ijls})_{ (i,j,l)\times s}$  has rank $m$, relation \eqref{2} implies that $c_1, \ldots, c_m \in \gamma_3(H)$. 	In particular, we have $c_1, \ldots, c_m \in \Phi(H)$ and hence $|H/\Phi(H)|\leq p^k$.
	 
	 
	  As $H$ has nilpotency class at most $3$, $\gamma_2(H)$ is abelian.   The group $\gamma_2(H)$ is generated by $\gamma_3(H)$ and elements of order~$p$ by relation \eqref{3}, and hence $\gamma_2(H)/\gamma_3(H)$  is elementary abelian. Thus so is $\gamma_2(H)/\gamma_2(H)\cap \ZZ(H)$.   On the other hand, the image of the map $H/\ZZ(H)\times \gamma_2(H)/\gamma_2(H) \cap \ZZ(H)\to \gamma_3(H)$  
	  induced by the commutator map generates $\gamma_3(H)$. For $h \in H$ and $h' \in \gamma_2(H)$, we have  
	$[h \ZZ(H), h' \gamma_2(H)\cap \ZZ(H)]^p= [h\ZZ (H), (h')^p \gamma_2(H)\cap \ZZ(H)]=1$ (see \Cref{lemma:computation}). Thus $\gamma_3(H)$ is elementary abelian.
	This implies that $\gamma_2(H)$ is elementary abelian (as an abelian group generated by elements of order~$p$).  %\ChDiego{[I implemented your proposals in this paragraph and in the following one. Plesase check. ]} \ChSofia{I fixed some typos, otherwise it's fine for me.}
	
 By relations \eqref{2} and \eqref{5}, we have $\gamma_3(H) = \langle c_1, \dots, c_m\rangle$, so
	\begin{equation}\label{eq:gamma3h}
 |\gamma_3(H)|\leq  p^m=|\gamma_3(\widetilde G)|.
	\end{equation} 
 Moreover, using \eqref{4}, we derive  that $|H/\ZZ(H)\Phi(H)|\leq p^d$. 
	Observe that $[H,\Phi(H)] \subseteq \gamma_3(H)$ due to $[x^p,y] \gamma_3(H) =  [x,y]^p \gamma_3(H)$ for all $x,y \in H$ (see Lemma~\ref{lemma:computation}) and $\gamma_2(H)$ being elementary abelian. Then the commutator in the group defines a bilinear map  $H/Z(H) \Phi(H) \times H/Z(H) \Phi(H)\to  \gamma_2(H)/\gamma_3(H)$, $(x\ZZ(H)\Phi(H), y\ZZ(H)\Phi(H))\mapsto [x,y] \gamma_3(H)$.
	As for $\widetilde{G}$, %\ChSofia{maybe "as for $\widetilde{G}$"`?}\ChDiego{[Done]} 
	we derive that
	\begin{equation}\label{eq:quotientgamma2gamma3}
	|\gamma_2(H)/\gamma_3(H)|\leq p^{{d\choose 2}}=|\gamma_2(\widetilde G)/\gamma_3(\widetilde G)|. 
	\end{equation}
	 Finally, $H/\gamma_2(H)=\GEN{g_1 \gamma_2(H),\dots, g_k \gamma_2(H)}$ has order at most $\prod_{i=1 }^k p^{n_i}=|\widetilde G/\gamma_2(\widetilde G)|$ by relation \eqref{1}. Since $ |\widetilde G|= |\widetilde G/\gamma_2(\widetilde G)|\cdot |\gamma_2(\widetilde G)/\gamma_3(\widetilde G) |\cdot  |\gamma_3(\widetilde G)| $, combining this with \eqref{eq:gamma3h} and~\eqref{eq:quotientgamma2gamma3} yields that $|H|\leq |\widetilde G|$. This proves the statement. 
\end{proof}

%\ChAngel{[Make clear when  $p>2$ is used]}\ChDiego{[I think one way or the other it is used everywhere. It would be nice though to have \Cref{corollary:1} also for $p=2$.]} \ChSofia{I thought about marking usage of $p \neq 2$ this morning, but I think it does not make sense (although I do see the point of the objection)}

% \ChDiego{[I changed the proof of the theorem and removed the useless lemmas. Can you check if it makes sense now?]}

%\ChSofia{at some point, we have to mention politely that there is a gap in Leo's proof. Otherwise, the next theorem seems redundant.}\ChDiego{[I don't think the theorem is redundant, even in if  Leo-Tobias's proof was completely correct. Observe that their result is not enough for us since our groups can be three-generated, and in contrast our corollary suffices. Moreover our theorem handles with groups with an arbitrary number of non-central generators of class three, which, I think, has only a precedent in \cite{Her07}. I will, however, mention this mistake to Leo when I see him.  ]}
%\ChSofia{of course, I forgot their assumption with the 2 generators when writing this comment}

%\ChDiego{[Should \Cref{theorem:1} be in the introduction as our second main result after \Cref{theorem:p3}? ]}
%\ChSofia{we can mention that as a key ingredient, although the conditions do maybe look a bit technical}
                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          
                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          
\begin{theorem}\label{theorem:1}  Let $p$ be an odd prime number and let $F$ be the field with $p$ elements. Consider finite $p$-groups $G$ and $H$ and           
	           %$G$ and $H$ be finite $p$-groups such that $F G\cong F H$, and 
	           let $|G/\Phi(G)\ZZ(G)|=p^d$. If 
	 %          \ChSofia{maybe rename those to (a) and (b) for consistency and because we refer to the numbers of the relations in 3.3}\ChDiego{[Done]}
	           \begin{enumerate}[(a)]
		\item\label{lemma:presentation1}   $G^p\cap \gamma_2(G)\subseteq \gamma_2(G)^p \gamma_3(G)$ and 
		\item\label{lemma:presentation2}  $|\gamma_2(G)/\gamma_2(G)^p\gamma_3(G)|=p^{{d\choose 2}} ,$
	\end{enumerate}
then $G/\gamma_2(G)^p \gamma_4(G)\cong  H/\gamma_2(H)^p \gamma_4(H)$.
\end{theorem}
\begin{proof}
%	\ChSofia{here, I reordered the arguments because this automatically explains the line starting with $p^d$. I am not sure, though, whether it should be $G$ or $\pi''(G)$ (because in the end, we are also studying $\pi''(H)$).} \ChDiego{[Ok to the reordering. I think   with $G$ is alright. ]}
Consider the isomorphism of elementary abelian $p$-groups 
%\ChSofia{reference/citation} \ChDiego{We might cite \cite[Proposition III. 1.14 and Proposition III.1.15]{Seh78}}
$$\varphi \colon G/\Phi(G)\to I(G)/I(G)^2, \quad x\Phi(G)\mapsto (x-1) + I(G)^2$$
(see~\cite[Propositions~III.1.14 and~III.1.15]{Seh78}). 
We have an $F$-vector space decomposition $\ZZ(FG) = F\ZZ(G) + \ZZ(FG) \cap K(FG)$ (see \cite[Lemma 6.10]{Sandling85}). It is well-known that $FG \cdot K(FG) = K(FG) \cdot FG = I(\gamma_2(G)) FG$. As $\gamma_2(G) \subseteq \Phi(G) = (1 + I(G)^2) \cap G$ holds (see \cite[Theorem~5.5]{Jen41}),
 we have 
 \begin{equation}\label{eq:kg}
K(FG) \subseteq I(G') FG \subseteq I(G)^2.
 \end{equation} In particular, $\varphi$ restricts to an isomorphism
\begin{equation}\label{eq:isocenters}
\ZZ(G)\Phi(G)/\Phi(G) \to  \left( \ZZ(FG)\cap I(G)+I(G)^2 \right)/I(G)^2.
\end{equation}	
Of course, the analogous reasoning applies to $H$. In particular, we obtain
$$p^d=|G/\Phi(G)\ZZ(G)|=|I(G)/(\ZZ(FG) \cap I(G) + I(G)^2)|= |H/\Phi(H)\ZZ(H)|.$$
Moreover, note that the conditions \eqref{lemma:presentation1} and \eqref{lemma:presentation2} depend only on the quotient $G/\gamma_2(G)^p \gamma_3(G)$, which is determined by $FG$ by \cite{San89}. Thus, if $G$ satisfies \eqref{lemma:presentation1} and \eqref{lemma:presentation2}, then the analogous statements hold for $H$. 
 
Consider the group $V$ and the projection $\pi \colon V(FG) \to V$ introduced in \eqref{eq:defpi}. Let $\{g_1,g_2,\dots, g_k, c_1,\dots,c_m\}$ be a set of generators of $\pi(G)$ satisfying relations $$\mathcal R_{( n_i,\alpha_{ij},\beta_{ijls})} ( g_1,\dots, g_k,c_1,\dots, c_m) $$ as in \Cref{lemma:presentation}. For $1\leq i \leq k$, the decomposition $\VSdd =\pi(H)\rtimes E$ (see \eqref{eq:decompv}) guarantees that $g_i=h_i e_i$ for some unique $h_i\in \pi(H)$ and $e_i\in E$. Moreover $c_j \in \gamma_2(\pi(G))=\gamma_2(S)=\gamma_2(\pi(H))$  for $j=1,\dots, m$  (see~\eqref{eq:gammaivs}). Now let $d<i \leq k$, so $g_i \in Z(\pi(G))$. We have 
\[(g_i -1) + I(G)^2 = (h_i e_i-1)+ I(G)^2 = h_i -1 + e_i - 1 + I(G)^2 = (h_i-1) + I(G)^2,\] 
using  $e_i -1 \in I(G)^2$ (see \cref{lemma:definitiona}).  
%$e_i -1 \in I(G') \subseteq I(G)^2$ (see \Cref{lemma:definitiona} and \eqref{eq:kg}). 
In particular, this yields $h_i -1 + I(G)^2 \in (Z(FG) \cap I(G) + I(G)^2)/I(G)^2$. Using the isomorphism in \eqref{eq:isocenters}, and the analogue for $H$, we derive that $h_i =\hat h_i \widetilde h_i$, with $\hat h_i\in \ZZ(\pi(H))
$ and $\widetilde h_i\in\Phi (\pi( H)) = \gamma_2(\pi(H)) \pi(H)^p$. Observe that $\pi(H)^p$ is central in $\pi(H)$ by \Cref{lemma:AgemoCentral}, so we can furthermore assume that $\widetilde h_i \in \gamma_2(\pi(H))$.
Then $\{h_1,\dots, h_d, \hat h_{d+1}, \dots , \hat h_k,c_1,\dots, c_m\}$ forms a generating set of $\pi(H)$. We now show that it satisfies the relations $$\mathcal R_{ (n_i,\alpha_{ij},\beta_{ijls})} ( h_1,\dots, h_d, \hat h_{d+1}, \dots , \hat h_k,c_1,\dots, c_m). $$  Then~$H$ is an epimorphic image of $G$ of the same size and we obtain $G\cong H$ as desired. 
	  
	  We check the relations one by one. Relations \eqref{3}, \eqref{4} and \eqref{5} are immediate.   
	  For $1\leq i \leq k$, one has $$g_i^p=(h_ie_i)^p=h_i^pe_i^p [h_i,e_i]^p=h_i^p,$$ and, in particular for $d<i\leq k$, one has $g_i^p=h_i^p= \hat h_i^p \widetilde h_i^p = \hat h_i ^p$. Then condition \eqref{1} follows. Condition \eqref{2} also follows readily because, for  $1\leq i \leq d$  and arbitrary $x_i,x_j,x_l\in  A$, we have that 
	  \begin{equation}
	  [[g_ix_i, g_jx_j], g_lx_l] =[[g_i,g_j] [x_j, g_j][g_i,x_j] , g_l x_l ] =[[g_i,g_j], g_l x_l]=[[g_i,g_j],g_l].
	\end{equation}
This completes the proof. 
\end{proof}


The next result shows that \Cref{theorem:1} is a generalization of \cite[Theorem 1.2]{MM20}. 
%\ChSofia{maybe I got something wrong, but in my version, the source does not contain a Proposition 2.5}\ChDiego{[I had a very old version. Now it should agree with the published version.]}
\begin{corollary}\label{corollary:1}
	Let $p$ be an odd prime number, and let $G$ and $H$ finite $p$-groups such that $F G\cong F H$. If $|G/\Phi(G)\ZZ(G)|=p^2$, then $$G/\gamma_2(G)^p \gamma_4(G)\cong H/\gamma_2(H)^p\gamma_4(H).$$ 
\end{corollary}
\begin{proof}
 If $G/ \gamma_2(G)^p \gamma_4(G)$ has nilpotency class $2$, then so does $H/\gamma_2(H)^p \gamma_4(H)$ (see \eqref{eq:gammaivs}) and hence the result follows by~\cite{San89}. Thus we assume otherwise. Observe that $\gamma_2(G)/\gamma_2(G)^p\gamma_3(G)$ has order $p$, as it is generated by $[g_1,g_2]\gamma_2(G)^p\gamma_3(G)$ if we consider a set of generators $\{g_1,\dots ,g_k,c_1,\dots, c_m\}$   as in \Cref{lemma:presentation}.  Observe that every $p$-power is central in $G/ \gamma_2(G)^p \gamma_4(G)$ since $[x,y^p] \equiv [x,y,y]^{{p\choose 2}} \equiv 1 \pmod{\gamma_2(H)^p \gamma_4(H)}$ for all $x,y \in G$ (by Lemma~\ref{lemma:computation} and using $p>2$). On the other hand, the image of $[g_1,g_2]$ in $G/\gamma_2(G)^p \gamma_4(G)$ is not central as~$G$ does not have nilpotency class $2$. This yields that $G^p \cap \gamma_2(G)\subseteq \gamma_2(G)^p \gamma_3(G)$. Thus the hypotheses of \Cref{theorem:1} with $d=2$ hold, and the statement follows. 
\end{proof} 


\section{\texorpdfstring{Groups with a center of index $p^3$}{Groups with a center of index p^3}}\label{sec:indexp3}

Let $p$ be an odd prime number. In this section, we use the results about the small group algebra from the previous section to give a positive answer to the modular isomorphism problem for groups with center of index $p^3$. 
%\ChSofia{we should opt for either $\gamma_2$ or $G'$ in the next proof}\ChDiego{[I think there are only $\gamma_2$ now. ]}
 
\begin{lemma}\label{lemma:descriptionp3}
%	Let $p> 2$ be a prime number and l
%\ChSofia{I know that I voted for not putting the $p > 2$ here, but here we need it, so should we maybe put it again? I am not sure.}\ChDiego{[I'm neutral]}
	Let $p$ be an odd prime number and let $G$ be a  finite $p$-group with $|G:\ZZ(G)|=p^3$. Then:
	\begin{enumerate}
		\item The nilpotency class of $G$ is at most $3$.
		\item $\gamma_2(G)$ is elementary abelian. 
		\item Either $G$ has nilpotency class $2$ or $|G/\Phi(G)\ZZ(G)|=p^2$.
	\end{enumerate}
	
\end{lemma}
\begin{proof}
Examination of the upper central series yields that the nilpotency class of $G$ is at most 3. The quotient $G/\ZZ(G)$ is isomorphic to one of the five groups of order $p^3$, namely to one of $C_{p^3}$, $C_p^3$, $C_{p^2}\times C_p$,  $(C_p\times C_p)\rtimes C_p$ or $C_{p^2}\rtimes C_p$. Observe that $G/Z(G)$ is not cyclic as $G$ is abelian otherwise.
	
 	Next, we show that $\gamma_2(G)$ is elementary abelian. Then $\gamma_2(G)$ is abelian, as $[\gamma_2(G), \gamma_2(G)]\subseteq \gamma_4(G)=1$. Thus it suffices to show that the generators of $\gamma_2(G)$ have order $p$. Note that $\gamma_3(G)$ is elementary abelian.  Indeed, it is generated by  elements of the form $[x,y]$ with $x\in G$ and $y\in \gamma_2(G)$; since $\gamma_2(G)\ZZ(G)/\ZZ(G)$ has order at most $p$, we have $y^p\in \ZZ(G)$, so \Cref{lemma:computation} shows that $1=[x,y^p]=[x,y]^p $. 
 	Let  $a,b_1, b_2\in G$ be such that $G/\ZZ(G)=\GEN{a\ZZ(G), b_1 \ZZ(G),   b_2\ZZ(G)}$. By the structure of $G/\ZZ(G)$  we may assume that $ b_i^p\in \ZZ(G) $ for $i=1,2$. Then for every $x\in G$ and $i \in \{1,2\}$, \Cref{lemma:computation} yields that
	$$1=[x , b_i^p]= [x,b_i]^p [[x,b_i],b_i]^{{p\choose 2}} =[x,b_i]^p.$$
	This shows that $\gamma_2(G)$ is elementary abelian. 
	
	Finally, if $G$ has nilpotency class $3$, then $G/\ZZ(G)$ is isomorphic to one of the two non-abelian $p$-groups of order $p^3$. Then it is clear that $|\gamma_2(G)\ZZ(G)/\ZZ(G)|=p$, and hence $|G/\Phi(G)\ZZ(G)|=p^2$.
\end{proof} 

We can now prove our main result:
 
	\begin{theorem}\label{theorem:p3}
 Let $p$ be an odd prime number, let $F$ be the field with $p$ elements,  and 
		let $G$ and $H$ be  finite $p$-groups. Suppose that $|G:\ZZ(G)|=p^3$. If $FG\cong FH$, then $G\cong  H$. 
	\end{theorem} 
		 \begin{proof}
		 	By \Cref{lemma:descriptionp3}, we know that $G'$ is elementary abelian. If the nilpotency class of $G$ is $2$, then $G\cong H/\gamma_2(H)^p \gamma_3(H)$ by \cite{San89}, so $G\cong H$ due to $|G|=|H|$. Thus, by \Cref{lemma:descriptionp3}, we can assume that the nilpotency class of $G$ is $3$, and that $|G/\Phi(G)\ZZ(G)|=p^2$. By \Cref{corollary:1}, we then have $G\cong  H$.
		 \end{proof}
	 
	 \begin{remark}
We point out that there cannot be an analogue of \Cref{theorem:p3} for $p=2$, as in \cite{GarciaMargolisdelRio} non-isomorphic finite $2$-groups with centers of index $8$ and isomorphic group algebras over every field of characteristic $2$ are presented. Hence our result underlines the difference between the cases $p=2$ and $p>2$ for this problem (we refer to \cite{GLdRS2022} for other contrasts).  
	 \end{remark}
	 
%	 \ChSofia{maybe the next remark should be presented earlier (if this was not the case, then the whole paper would be meaningless)} \ChDiego{[I am not sure if these examples deserve all the trouble to check that they are non-isomorphic. If they stay, I'm okay with presenting this earlier. ]}
%	 

We conclude this section with examples illustrating that the class of $p$-groups with a center of index $p^3$ is not contained in one of the classes for which the modular isomorphism problem was previously known to have a positive answer. 
%We begin with some concrete examples:
%\ChDiego{[I think it is nicer to present the examples in pairs, so the reader can indeed be sure that but ours no known result allows to differentiate their group algebras. To me a group alone is not as meaningful. It might happen, as happens with your group  $G = \texttt{SmallGroup}(5^7,30087) $, that no other group is a candidate to be an alternative group basis for $kG$, which can be checked using the GAP method MIPBinsByGT in ModIsomExt.  I attach the list of bins for the groups in your list.]} \ChSofia{I agree. I didn't check with ModIsomExt, but of course this should be done.}
%
%
%\ChDiego{[I commented the old example and added a new one, using one of the resulting bins.]}
\begin{example}
$\null$
\begin{enumerate} 
 	\item Consider the following groups of order $5^7$: 
%\ChDiego{[Not sure about the big angles. We can delete   one of the groups if you prefer so.]} \ChSofia{I would include them all and I like the big angles. Maybe it makes it easier to read if one has some space between the groups, but I am not sure. }
	\begin{align*}
	G&=\GEN{a,b,c,d  \left| \begin{array}{c} 
			a^{625} = b^5=c^{25}=[b,a]^5=[c,a]=[c,b] =[d,a]=[d,b] =[d,c] =1, \\  a^5=d, 
			\left[  [b,a], a \right] = d^5, [[b,a],b]= c^5    	\end{array} \right. } \\ 
		&\\ 
		H&=    \GEN{a,b,c,d  \left| \begin{array}{c} 
				a^{625} = b^5=c^{25}=[b,a]^5=[c,a]=[c,b] =[d,a]=[d,b] =[d,c] =1, \\  a^5=d, 
				\left[  [b,a], a \right] = c^5, [[b,a],b]= d^5    	\end{array} \right. } \\
			&\\
			K&=	    \GEN{a,b,c,d  \left| \begin{array}{c} 
					a^{625} = b^5=c^{25}=[b,a]^5=[c,a]=[c,b] =[d,a]=[d,b] =[d,c] =1, \\  a^5=d^2, 
					\left[  [b,a], a \right] = c^5, [[b,a],b]= d^5    	\end{array} \right. }. 
	\end{align*} 
In {\sf GAP}, they are listed as ${ \rm \tt SmallGroup}(5^7, 1599)$, ${\rm \tt SmallGroup}(5^7,1734)$ and ${\rm \tt SmallGroup}(5^7,1766)$, respectively.
%	that is,  
%	\begin{align*}
%		G &= \texttt{SmallGroup}(5^7, 1599),  \\
%		  H&= \texttt{SmallGroup}(5^7, 1734), \\
%		 K&= \texttt{SmallGroup}(5^7, 1766 ).
%	\end{align*} 
	 Each of these groups is $3$-generated with minimal set of generators $\{a,b,c\}$. The center is given by $ \langle c,d \rangle \cong C_{25} \times C_{25}$, and hence it has index $p^3$, so Theorem~\ref{theo:maintheorema} applies to $G,H$ and $K$. On the other hand, a quick verification shows that none of these groups is   covered by any of the cases for which MIP is known (according to the survey in \cite{Mar22}). Moreover, these three groups agree in all  the group theoretical invariants implemented in the {\sf GAP} package ModIsomExt~\cite{GAP, ModIsomExt}.
	 %a,b,c,d correspond to f1, f2, f3 , f5 
	 
%	\item Let $G = \texttt{SmallGroup}(5^7,30087) \cong (C_{25} \times ((C_{25} \rtimes C_5) \rtimes C_5)) \rtimes C_5$
%	 given by the presentation 
%	\begin{alignat*}{1}
%	G = \langle a,b,c, d \mid &a^5 = b^5 = c^{25} = d^{25} = [a,b]^5 =[a,c] = [b,c] = [c,d] = [a,d] = [b,d] = 1,\\
%	 &[[a,b], a] = c^{5}, [[a,b], b] = d^{5}\rangle .
%	\end{alignat*}
%%a,b,c,d correspond to f1, f2, f3^(-1), f4^(-1) in GAP 
%%(C25 x ((C25 : C5) : C5)) : C5"
%
%We have $\ZZ(G) = \langle c,d \rangle \cong C_{25} \times C_{25}$ and hence $|G: \ZZ(G)| = 5^3$, so Theorem~\ref{theo:maintheorema} applies to $G$. On the other hand, a quick verification shows that $G$ is not covered by any of the cases for which MIP is known (according to the survey in \cite{Mar22}). 

%list of SmallGroup-indices of order $5^7$ with the following properties:
% class 3, center of index 5^3, more than 2 generators, $C_G(G')$ not a maximal subgroup of $G$ (it is always abelian)
%[ 1491, 1492, 1493, 1494, 1495, 1496, 1497, 1498, 1499, 1599, 1600, 1601, 
%1602, 1603, 1604, 1731, 1732, 1733, 1734, 1763, 1764, 1765, 1766, 1848, 
%1849, 1850, 1851, 1852, 1853, 1854, 1855, 1856, 1857, 1858, 1859, 1972, 
%1978, 1986, 2008, 30087, 30088, 30089, 30090, 30091, 30092, 30093, 30094, 
%30095, 30096, 30097, 30098, 30099, 30100, 30101, 30102, 30103 ]
	
	
% groups of order 7^6: 
% need to have nilpotency class > 2 since G' has exponent p (and this is covered by Margolis). 
% class 3, not 2-generated, C_G(G') not maximal, no cyclic subgroup of index 7^5: [ 135, 138, 140, 141, 152, 154, 155, 156, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169 ] (18 groups)


\item 	
Let $p \geq 7$. In the list of groups of order $p^6$ given in \cite{NEW23}, consider the groups $G_{6,3}$ as well as $G_{6,5r}$ for $r \in \{1, \nu\}$. These groups have centers of index $p^3$ and do not lie in the union of classes of $p$-groups for which the modular isomorphism problem is known to have a positive answer. For small values of~$p$, one can verify computationally that these groups agree in all group theoretical invariants implemented in ModIsomExt. 
%\ChSofia{I was thinking that we should maybe comment more on why $p^6$ is interesting and why we chose groups of order $5^7$}
%\ChDiego{[I'm not sure. Maybe we should mention in the introduction what is known for groups of small order? I wrote  something in this direction there.]}
% In all three cases, the isomorphism between the groups in GAP and the presentation of O'Brien is given by: a_1 \mapsto \alpha_4, a_2 \mapsto \alpha_5, a_3 \mapsto \beta_1, a_4 \mapsto \alpha_3^{-1}, a_5 \mapsto \alpha_1^{-1}, a_6 \mapsto \alpha_2^{-1}





%
%
%Out of the 43 isoclinism families of groups of order $p^6$, the groups in the classes $\Phi_3$, $\Phi_4$, $\Phi_6$ and $\Phi_{11}$ have centers of index $p^3$. For instance, for $p = 7$, we find $18$ groups of order $7^6$ and centers of index $7^3$ for which MIP was not known to have a positive answer previously (according to \cite{Mar22}). 
% \ChSofia{I am referring to O'Brien's list here, need to check the other conditions. I hope that we get several groups here, otherwise this would be unimpressive.} \ChDiego{[For $p^7$ there are surely examples. For $p^6$ seems possible but less likely: we need at least three generators, and the commutator has order at least $p^2$ to obtain nilpotency class $3$. So interesting groups have order at least  $p^5$. If wou want some freedom in the orders of the generators,  the group becomes too big. ]} \ChSofia{the problem is that there is no list of groups of order $p^7$ (as far as I know), so this only works for $p^6$. I'll check the ones for $p^6$ and if they don't work, we leave out this item and just do a concrete example?} \ChDiego{[According to my calculations, once removed the groups with abelian direct factors and after applying MIPBinsByGT, only three groups remain. See the .txt. It is not as impressive as $18$, but at least there are examples (I suppose that for each $p\geq 7$ there will be at least $3$ examples of order $p^6$ in these conditions, if not more).]}
\end{enumerate}
\end{example}

%\ChDiego{[Once we have the previous examples, I think we should delete this. Specially if we manage to include examples of orden $p^6$ for each $p$.]}A more involved example is given by the following: 
%	 
%	 \begin{remark}
%	 	Observe that the class of $p$-groups $G$ with $|G:\ZZ(G)| = p^3$ is not contained in the union of the classes of $p$-groups for which the modular isomorphism problem is known to have a positive answer. Indeed, let $p$ be an odd prime,  $m\geq 3$ and $m,l\geq 1$ and $m>n$. For $1\leq u \leq p-1$, consider the group  $G_u$ generated by $x,y,z$ and defined by the following relations: 
%	 	\begin{enumerate}
%	 		\item   $x^{p^{m}}= [[y,x],y]^u$. 
%	 		\item $ y^{p^n}=1$,
%	 		\item $ z^{p^l}=[[y,x], x]$, 
%	 		\item $G_u'$ elementary abelian. 
%	 		\item Nilpotency class $3$. 
%	 		\item $[z,x]=[z,y]=1$. 
%	 	\end{enumerate}
%	 	It is clear that $|G_u:\ZZ(G_u)|=p^3$, and we will show below that if $u$ is an square modulo $p$ but $v$ is not, then $G_u\not \cong G_v$. These groups do not lie in the union of the classes of $p$-group for which the modular isomorphism problem is known to have a positive answer (according to the recent survey \cite[Section 3]{Mar22}). Moreover, for $p=3$, $n=4$ and $l=m=1$, we are able to check computationally that all the group theoretical invariants implemented in the {\sf GAP} package ModIsomExt~\cite{GAP, ModIsomExt} agree for $G_1$ and $G_2$. 
	 	
	 	
%	 	Now we show that if $u$ is an square modulo $p$ and $v$ is not, then $G_u\not \cong G_v$.To simplify notation write $G=G_u$. Observe that $$G/G' \ZZ(G)=\GEN{xG'\ZZ(G), y G'\ZZ(G)}\cong C_p\times C_p\qand  \Omega_n(G)G'\ZZ(G)/G' \ZZ(G)=\GEN{y G' \ZZ(G)}\cong C_p.$$  Let $$\mathfrak c: G/G'\ZZ(G)\times \Omega_n(G)G'\ZZ(G)/G' \ZZ(G) \to \gamma_3(G), \quad (aG'\ZZ(G), b G'\ZZ(G)) \mapsto [[b,a],a].  $$It is easy to see that this map is well defined and
%	 	$$\mathfrak c(x^{\alpha} y^{\varepsilon}G'\ZZ(G),   y ^\beta G' \ZZ(G))=[[y,x],y]^{\alpha \beta^2 }.$$ Denote also $\Lambda: G/G'\ZZ(G)\to \mho_m(G)$ be the $p^m$-power map $xG'\ZZ(G)\mapsto x^{p^m}$.   Since $\exp(G )=p^{m+1}$, we have that $\exp(G'\ZZ(G))=p^{m }$ and hence  $\Lambda$ is well-defined. It suffices to show the following equivalence: there is an element $b\in  \Omega_n(G)G'\ZZ(G)/G' \ZZ(G)$ such that $\Lambda(a )=\mathfrak c (a,b)$ for each $a \in G/G'\ZZ(G)$ if and only if $u$ is a square modulo $p$. Indeed, $ \mathfrak c(x^{\alpha} y^{\varepsilon}G'\ZZ(G),   y ^\beta G' \ZZ(G))=\Lambda(x^\alpha y ^{\varepsilon} G' \ZZ(G)) $
%	 	if and only if $[[y,x],y]^{\beta^2-u}=1$  if and only if $\beta ^2\equiv u \mod p$.
%	 \end{remark}
 
 

%\section{Groups with center of index $p^4$ and class $2$}
%	  
%	  
%	  \begin{lemma}\label{lemma:descriptionp4c2}
%	  	Let $G$ be finite $p$-group with nilpotency class $2$ such that $|G:\ZZ(G)|=p^4$. Then one of the the following holds:
%	  	\begin{enumerate}
%	  		\item $G'$ is elementary abelian. 
%	  		\item $G/\ZZ(G)\cong C_{p^2}\times C_{p^2}$, and $G'$ is cyclic of order   $p^2$.
%	  	\end{enumerate}
%	  \end{lemma}
%  \begin{proof}
%  	  As $G$ has nilpotency class $2$, $\gamma_2(G)$ is abelian and $G/\ZZ(G)$ is an abelian group of order $p^4$. Suppose first that it is not isomorphic to $C_{p^2}\times C_{p^2}$. Then it has a Burnside basis $x_1,\dots, x_k$ such that $x_i^p=1$ for $i>1$. Since the commutator map $G/\ZZ(G)\times G/\ZZ(G)\to \gamma_2(G)$ is bilineal and surjective, it follows that $\gamma_2(G)$ is elementary abelian. Now suppose that $G/\ZZ(G)$ is isomorphic to $C_{p^2}\times C_{p^2}$. Then another immediate argument with the bilineal commutator map shows that shows that $\gamma_2(G)$  is cyclic of order dividing $p^2$.
%    \end{proof}
%
%
%By \cite{San89}, the modular isomorphism problem has positive answer for the groups satisfying the first condition in \Cref{lemma:descriptionp4c2}. Thus we only need to consider groups satisfying the second condition.  
%
% 
%
%\ChDiego{[I commented the rest, as at this moment a real proof seems out of our reach.]}

\section*{Acknowledgments} 
The authors are grateful to Leo Margolis and Ángel del Río for their comments on an earlier version of this paper.



%\ChDiego{[By the previous lemma and \cite{San89}, the only interesting case is when $G/\ZZ(G)\cong C_{p^2}\times C_{p^2}$ and $G'$ is cyclic of order $p^2$. This should be easy to attack directly. However, I wanted to do something more general, like \Cref{proposition:BdR}. But I  think the tentative \Cref{lemma:descriptionBdR}  is not correct. Maybe we could tray to find a correct version, deal with the $|G'|=p^2$ version directly. Either way we can obtain \Cref{theorem:p4c2}]}
%
%To deal with the second case in the previous lemma we use \Cref{proposition:BdR}, which generalizes the main result in \cite{BdR20} for $p>2$, and is based on the following lemma. 
%
%\begin{lemma}\label{lemma:descriptionBdR}
%	Let $G$ and $H$ indecomposable finite $p$-groups of nilpotency class $2$ such that $|G/\Phi(G)\ZZ(G)|=p^2=|H/\Phi(H)\ZZ(H)|$. Then $G'$ is cyclic. Moreover,   $G\cong H$ if and only if  the following conditions hold:
%	\begin{enumerate}
%		\item $G/\Omega_1(G')\cong H/\Omega_1(H')$,
%		\item $\exp(G)=\exp(H)$, and
%		\item $\ZZ(G)\cong \ZZ(H)$.
%	\end{enumerate}
%	\end{lemma}
%\begin{proof}
%	Let $G$ and $H$ be finite $p$-groups as in the statement. Let $x_1,\dots, x_k$ be a Burnside basis of $G$ such that $x_i\in \ZZ(G)$ for each $i>2$. Write $G_0=\GEN{x_1,x_2}$. Then $\gamma_2(G)=\gamma_2(G_0)$ is the derived subgroup of a $2$-generated group of nilpotency class $2$, and hence cyclic. Denote $z=[x_2,x_1]$, so $\gamma_2(G)=\GEN{z}$.   
%	
%	TO BE COMPLETED.
%\end{proof}
%\begin{proposition}\label{proposition:BdR}
%	Let $G$ and $H$ be a finite $p$-groups. Suppose that $G$ is of nilpotency class $2$ such that $|G/\Phi(G)\ZZ(G)|=p^2$. If $FG\cong FH$, then $G\cong H$.
%\end{proposition}
%  \begin{proof}
%  	By \Cref{lemma:descriptionBdR}, $G'$ is cyclic.
%  	  	We argue by induction on $|G'|$. If $|G'|=p$, then result follows from \cite{San89}. Suppose the proposition holds groups with derived subgroup of order $p^n$, $n\geq 1$, and that $|G'|=p^{n+1}$. Since $G'$ is cyclic, if $G$ is decomposable then $G=G_1\times A_G$, where $A_G$ is abelian and $G_1$ is indecomposable with $G_1'=G'$. Moreover, by the main result of \cite{GL2022} $FG\cong FH$ implies that $H=H_1\times A_H$, where $A_G\cong A_H$ and $H_1$ is indecomposable, and $F G_1\cong F H_1$. In particular $H_1'=H' $. Thus we can assume that $G$ is indecomposable.  Let $\phi:FG\to FH$ be an isomorphism preserving the augmentation. Recall $I(G')FG$ is the ideal of $FG$ generated by the space of Lie commutators $[FG,FG]$, so $\phi(I(G')FG)=I(H')FH$. As $G'$ is cyclic then $(I(G')FG)^{|G'|/p}= I(\Omega_1(G'))FG$, and the analogous holds for $H$. Thus $\phi(I(\Omega_1(G))FG)=I(\Omega_1(H))FH$. Hence $\phi$ induces an isomorphism $\bar \phi : F(G/\Omega_1(G'))\to F(H/\Omega_1(H'))$. By the induction hypothesis $G/\Omega_1(G')\cong H/\Omega_1(H')$, so, by \Cref{lemma:descriptionBdR}, $G\cong H$.
%  \end{proof}
%
%\begin{theorem}\label{theorem:p4c2}
%	Let $G$ and $H$ be  finite $p$-groups. Suppose that $G$ has class $2$ and $|G:\ZZ(G)|=p^4$. If $FG\cong FH$ then $G\cong H$. 
%\end{theorem}
%\begin{proof}
%	By \Cref{lemma:descriptionp4c2}, either $G'$ is elementary abelian, or $G/\ZZ(G)\cong C_{p^2}\times C_{p^2}$ and $G'$ is cyclic. If the former holds, then the result follows from \cite{San89}. So we suppose the latter. Then $|G/\Phi(G)\ZZ(G)|= p^2$, so the theorem follows from \Cref{proposition:BdR}.
%\end{proof}
%\newpage
%For a prime $p$ and a positive integer $n$, we denote $v_p(n)$ the largest non-negative integer $k$ such that $p^k$ divides $n$.  We use the convention $v_p(0)=+\infty$.  Recall that $p>2$
%\begin{lemma}Let   $G$ be a finite  $p$-group with   nilpotency class $2$ such that $G/\ZZ(G)$ is $2$-generated. Then  there are positive integers $m$ and $n_i$ and integers $0\leq \alpha_i,\beta_i,\delta_i <p^{n_i}$ such that $n_i\leq v_p(\delta_i)+m$ satisfying
%	$$ G\cong \GEN{ x,y,z_1,\dots ,z_n\mid \mathcal R_{n_i,\alpha_i,\beta_i,\delta_i}(x,y,z_1,\dots,z_k),}$$
%	where $R_{n_i,\alpha_i,\beta_i,\delta_i}(x,y,z_1,\dots,z_k)$ is the set formed by the following relations:
%	\begin{enumerate}[(a)]
%		\item $z_1,\dots, z_n\in \ZZ(G)$.
%		\item $z_i^{p^{n_i}}=1$ for $i\in\{1,\dots, k\}$. 
%		\item $x^{p^m}=z_1^{\alpha_1}\dots z_k^{\alpha_k}$.
%		\item $y^{p^m}=z_1^{\beta_1}\dots z_k^{\beta_k}$. 
%		\item $[y,x]= z_1^{\delta_1}\dots z_k^{\delta_k}$
%	\end{enumerate} 
%Moreover $G' \cong C_{p^m}$ and $G/\ZZ(G)\cong C_{p^m}\times C_{p^m}$. 
%\end{lemma}
%
%\begin{proof}
%	????
%\end{proof}
%
%
%For a positive integers $n$ and elements $x_1,\dots, x_n\in  FG$,  we define recursively $ [x_1,\dots, x_{n+1}]= [[x_1,\dots, x_n], x_{n+1}] $. In other words, we consider Lie commutators nested to the right. 
%\begin{lemma}
%	Let $G$ be a finite $p$-group of nilpotency class $2$ and $n$ a positive integer. Let $x, x_1,\dots, x_n \in I(G)$ and $y\in  I(G')^{n}FG$.  Then
%	\begin{itemize}
%		\item 	$[x,y]\in I(G')^{n+1}IFG$.
%		\item $[x_1,x_2,\dots, x_n]\in I(G')^nFG$. 
%	\end{itemize} 
%\begin{proof}
%	Considering $FG$ as a Lie algebra with the \cite[Chapter V, Section 7]{Jacobson}
%\end{proof}
% 
%\end{lemma}
%
%\begin{lemma}
%	Let $G$ be a finite $p$-group with nilpotency class $2$ such that $G'$ is cyclic of order $p^m$. If  $\alpha,\beta \in I(G)$, then 
%	$$(\alpha +\beta)^{p^m}= \alpha^{p^m}+\beta^{p^m} $$
%\end{lemma}
%\begin{proof}
%	contenidos...
%\end{proof} 
%As a consequence of this lemma, $(\alpha+\beta)^{p^{m-1}}\equiv \alpha^{p^{m-1}}+\beta^{p^{m-1}} \mod I(\gamma_2(G)^{p^{m-1}})FG$.
%\begin{proposition}
%	Let $p$ an odd prime and let $G$ be a finite  $p$-group with   nilpotency class $2$ such that $G/\ZZ(G)$ is $2$-generated. If $H$ is a group such that $FG\cong FH$, then $G\cong H$. 
%\end{proposition}
%\begin{proof}
%	contenidos...
%\end{proof}
%\begin{lemma}
%	Let $G$ be a finite $p$-group such that $G/\ZZ(G)\cong C_{p^2}\times C_{p^2}$. Then $G$ admits a presentation with generators $x,y,z$ and the following relations:
%	\begin{itemize}
%	 \item Nilpotency class $2$. 
%	 \item $[y,x]^{p^2}=1$. 
%	 \item $z$ is central
%	\end{itemize}
% 
%\end{lemma}
%
%\newpage 
%  \section{Groups with center of index $p^4$ and class $3$}
%  	  
%  \begin{lemma}\label{lemma:descriptionp4c3}
%  	Let $G$ be finite $p$-group with nilpotency class $3$ such that $|G:\ZZ(G)|=p^4$. Then one of the the following holds:
%  	\begin{enumerate}
%  		\item $G'$ is elementary abelian. 
%  		\item $G/\ZZ(G)\cong C_{p^2}\rtimes C_{p^2}$, and $G'\cong C_{p^2}\times C_p^m$, for some $m\geq 0$.
%  	\end{enumerate}
%  \end{lemma}
%  TO BE DONE
%  
%  \section{Groups with center of index $p^4$ and class $4$}
%  TO BE DONE
 \bibliographystyle{amsalpha}
\bibliography{MIP}

\end{document}

\providecommand{\bysame}{\leavevmode\hbox to3em{\hrulefill}\thinspace}
\providecommand{\MR}{\relax\ifhmode\unskip\space\fi MR }
% \MRhref is called by the amsart/book/proc definition of \MR.
\providecommand{\MRhref}[2]{%
  \href{http://www.ams.org/mathscinet-getitem?mr=#1}{#2}
}
\providecommand{\href}[2]{#2}
\begin{thebibliography}{GLMdR22}

\bibitem[Bag88]{BaginskiMetacyclic}
C.~Bagi\'{n}ski, \emph{The isomorphism question for modular group algebras of
  metacyclic {$p$}-groups}, Proc. Amer. Math. Soc. \textbf{104} (1988), no.~1,
  39--42.

\bibitem[Bag99]{Bag99}
\bysame, \emph{On the isomorphism problem for modular group algebras of
  elementary abelian-by-cyclic {$p$}-groups}, Colloq. Math. \textbf{82} (1999),
  no.~1, 125--136.

\bibitem[BK19]{BK19}
C.~Bagi\'{n}ski and J.~Kurdics, \emph{The modular group algebras of
  {$p$}-groups of maximal class {II}}, Comm. Algebra \textbf{47} (2019), no.~2,
  761--771.

\bibitem[Des56]{Deskins1956}
W.~E. Deskins, \emph{Finite {A}belian groups with isomorphic group algebras},
  Duke Math. J. \textbf{23} (1956), 35--40.

\bibitem[Dre89]{Drensky}
V.~Drensky, \emph{The isomorphism problem for modular group algebras of groups
  with large centres}, Representation theory, group rings, and coding theory,
  Contemp. Math., vol.~93, Amer. Math. Soc., Providence, RI, 1989,
  pp.~145--153.

\bibitem[Eic08]{Eick08}
B.~Eick, \emph{Computing automorphism groups and testing isomorphisms for
  modular group algebras}, J. Algebra \textbf{320} (2008), no.~11, 3895--3910.

\bibitem[GAP19]{GAP}
The GAP~Group, \emph{{GAP -- Groups, Algorithms, and Programming, Version
  4.10.2}}, 2019, \url{https://www.gap-system.org}.

\bibitem[GLdRS22]{GLdRS2022}
D.~Garc\'{\i}a-Lucas, \'{A}. del R\'{\i}o, and M.~Stanojkovski, \emph{{On Group
  Invariants Determined by Modular Group Algebras: Even Versus Odd
  Characteristic}}, Algebr. Represent. Theory (2022),
  \url{https://doi.org/10.1007/s10468-022-10182-x}.

\bibitem[GLMdR22]{GarciaMargolisdelRio}
D.~García-Lucas, L.~Margolis, and Á. del Río, \emph{Non-isomorphic
  $2$-groups with isomorphic modular group algebras}, J. Reine Angew. Math.
  \textbf{154} (2022), no.~783, 269--274.

\bibitem[HS07]{HertweckSoriano07}
M.~Hertweck and M.~Soriano, \emph{Parametrization of central {F}rattini
  extensions and isomorphisms of small group rings}, Israel J. Math.
  \textbf{157} (2007), 63--102.

\bibitem[Jen41]{Jen41}
S.~A. Jennings, \emph{The structure of the group ring of a {$p$}-group over a
  modular field}, Trans. Amer. Math. Soc. \textbf{50} (1941), 175--185.

\bibitem[{Mar}22]{Mar22}
L.~{Margolis}, \emph{{The Modular Isomorphism Problem: A Survey}}, Jahresber.
  Dtsch. Math. Ver. \textbf{124} (2022), 157--196.

\bibitem[MM20]{ModIsomExt}
L.~Margolis and T.~Moede, \emph{{ModIsomExt}: a {GAP 4} package, version
  1.0.0}, 2020, \url{https://www. tu-braunschweig.de/en/iaa/personal/moede}.

\bibitem[MM22]{MM20}
L.~Margolis and T.~Moede, \emph{The {M}odular {I}somorphism {P}roblem for small
  groups – revisiting {E}ick's algorithm}, Journal of Computational Algebra
  \textbf{1-2} (2022), 100001.

\bibitem[MS22]{MS22}
L.~Margolis and M.~Stanojkovski, \emph{On the modular isomorphism problem for
  groups of class $3$ and obelisks}, J. Group Theory \textbf{25} (2022), no.~1,
  163--206.

\bibitem[NOVL]{NEW23}
M.~F. Newman, E.~A. O'Brien, and M.~R. Vaughan-Lee, \emph{Presentations for the
  groups of order $p^6$ for prime $p\geq 7$},
  \url{https://arxiv.org/abs/2302.02677}.

\bibitem[Pas65]{Passman1965p4}
D.~S. Passman, \emph{The group algebras of groups of order {$p^{4}$} over a
  modular field}, Michigan Math. J. \textbf{12} (1965), 405--415.

\bibitem[Sal93]{SalimTesis}
M.~A.~M. Salim, \emph{The isomorphism problem for the modular group algebras of
  groups of order $p^5$}, ProQuest LLC, Ann Arbor, MI, 1993, Thesis
  (Ph.D.)--University of Manchester.

\bibitem[San85]{Sandling85}
R.~Sandling, \emph{The isomorphism problem for group rings: a survey}, Orders
  and their applications ({O}berwolfach, 1984), Lecture Notes in Math., vol.
  1142, Springer, Berlin, 1985, pp.~256--288.

\bibitem[San89]{San89}
\bysame, \emph{The modular group algebra of a central-elementary-by-abelian
  {$p$}-group}, Arch. Math. (Basel) \textbf{52} (1989), no.~1, 22--27.

\bibitem[San96]{San96}
\bysame, \emph{The modular group algebra problem for metacyclic {$p$}-groups},
  Proc. Amer. Math. Soc. \textbf{124} (1996), no.~5, 1347--1350.

\bibitem[Seh78]{Seh78}
S.~K. Sehgal, \emph{Topics in group rings}, Monographs and Textbooks in Pure
  and Applied Math., vol.~50, Marcel Dekker, Inc., New York, 1978.

\bibitem[SS95]{SalimSandling1995}
M.~A.~M. Salim and R.~Sandling, \emph{The unit group of the modular small group
  algebra}, Math. J. Okayama Univ. \textbf{37} (1995), 15--25.

\bibitem[SS96]{SalimSandlingp5}
\bysame, \emph{The modular group algebra problem for groups of order {$p^5$}},
  J. Austral. Math. Soc. Ser. A \textbf{61} (1996), no.~2, 229--237.

\bibitem[Whi68]{Whitcomb}
A.~Whitcomb, \emph{The {G}roup {R}ing {P}roblem}, ProQuest LLC, Ann Arbor, MI,
  1968, Thesis (Ph.D.) -- The University of Chicago.

\end{thebibliography}