Boolean-Cayley-graphs / papers-talks / Bent-functions-Cayley-graphs / Leopardi-Bent-functions-Cayley-graphs.body.tex
22144 views\section{Introduction}1\label{sec-Introduction}2Binary bent functions are important combinatorial objects.3Besides the well-known application of bent functions and their generalizations to cryptography4\cite{Ada97} \cite[4.1-4.6]{Tok15bent},5bent functions have well-studied connections to Hadamard difference sets \cite{Dil74},6symmetric designs with the symmetric difference property \cite{DilS87block,Kan75symplectic},7projective two-weight codes \cite{CalK1986,Din2015} and strongly regular graphs.89In two papers, Bernasconi and Codenotti \cite{BerC99}, and then Bernasconi, Codenotti and Vanderkam10\cite{BerCV01} explored some of the connections11between bent functions and strongly regular graphs.12While these papers established that the Cayley graph of a binary bent function (whose value at 0 is130) is a strongly regular graph14with certain parameters, they leave open the question of which strongly regular graphs with these15parameters are so obtained.1617In a recent paper \cite{Leo17Hurwitz},18the author found an example of two infinite series of bent functions whose19Cayley graphs have the same strongly regular parameters at each dimension,20but are not isomorphic if the dimension is 8 or more.2122Kantor, in 1983 \cite{Kan83exponential}, showed that the numbers of non-isomorphic projective linear23two weight codes with certain parameters,24Hadamard difference sets, and symmetric designs with certain properties, grow at least exponentially25with dimension.26This result suggests that the number of strongly regular graphs obtained as Cayley graphs of bent27functions also increases at least exponentially with dimension.2829%See also the paper by Jungnickel and Tonchev \cite{JunT1991}.3031A 2003 paper by Cameron \cite{Cam2003} considers random strongly regular graphs with given parameters,32and outlines some prerequisites for a theory of random strongly regular graphs, including that33``the number of non-isomorphic graphs is superexponential in the number of vertices.''3435The goal of the current paper is to further explore the connections between bent functions, their36Cayley graphs, and related combinatorial objects,37and in particular to examine the relationship between various equivalence classes of bent38functions, in particular, the relationship between the extended affine39equivalence classes and equivalence classes defined by isomorphism of Cayley graphs.40As well as a theoretical study of bent functions of all dimensions, an computational study is conducted41into bent functions of dimension at most 8,42using SageMath \cite{SageMath7517} and CoCalc \cite{CoCalc}.4344The methodology for this study is modelled on \emph{experimental mathematics}.45As stated by Borwein and Devlin \cite[pp. 4-5]{BorD2008}:46\begin{quotation}47What makes experimental mathematics different (as an enterprise)48from the classical conception and practice of mathematics is that the49experimental process is regarded not as a precursor to a proof, to be50relegated to private notebooks and perhaps studied for historical51purposes only after a proof has been obtained. Rather, experimentation52is viewed as a significant part of mathematics in its own right, to53be published, considered by others, and (of particular importance)54\emph{contributing to our overall mathematical knowledge.}55\end{quotation}5657The theoretical results listed this paper serve a few purposes.58First, in order to classify bent functions by their Cayley graphs,59it helps to understand the relationship between Cayley equivalence and other concepts of equivalence60of bent functions, especially if this helps to cut down the search space needed for the61classification.62A similar consideration applies to the duals of bent functions.63Second, some of the empirical observations made in the classification of bent functions in small64dimensions can be explained by these theoretical results.65Third, these theoretical results can improve our understanding of the relationships between66bent functions, projective two-weight codes, strongly regular graphs, and67symmetric block designs with the symmetric difference property.68In what follows, known results with references are presented as propositions, or sometimes as remarks;69and results where the statement or the proof seems to be missing or obscure within the existing literature70are presented as lemmas or theorems, with proofs.7172This paper makes no pretence at being an exhaustive survey, neither should it be construed that73the lemmas and theorems listed here make any claim to originality.74Even given the current excellent electronic search capabilities available, it would be folly to do so given75the extensive literature that has been generated on the study of bent functions since the 1960s.76Some recent surveys include books by Cusick and Stanica \cite{CusS2017},77Mesnager \cite{Mes2016} and Tokareva \cite{Tok15bent},78and the article by Carlet and Mesnager \cite{CarM2016four}.7980The remainder of the paper is organized as follows.81Section~\ref{sec-Preliminaries} covers the concepts, definitions and known results used later in the paper.82Section~\ref{sec-Bent-graphs} discusses the relationship between bent functions and strongly regular graphs.83Section~\ref{sec-Equivalence} introduces various concepts of equivalence of bent functions.84Section~\ref{sec-Bent-designs} discusses the relationship between bent functions and block designs.85%Section \ref{sec-Results} contains the main theoretical results of the paper.86Section~\ref{sec-Code} describes the SageMath and CoCalc code that has been used to obtain87the computational results of this paper.88Section~\ref{sec-Discussion} puts the results of this paper in the context of questions that are still open.89The appendices contain the proof of one of the properties of quadratic bent functions,90and list some of the properties of the equivalence classes of bent functions for dimension up to 8.9192\section{Preliminaries}93\label{sec-Preliminaries}9495% \begin{frame}96This section presents some of the key concepts used in the remainder of the paper.97We first examine Boolean functions, then define bent Boolean functions,98and finally explore the relationships between bent functions99and Hamming weights.100101% \begin{frame}102\subsection{Boolean functions}103\label{sec-Boolean-functions}104105Here and in the remainder of the paper, $\F_2$ denotes the field of two elements,106also known as $GF(2)$. Models of $\F_2$ include integer arithmetic modulo 2107($\Z/2\Z$ also known as $\Z_2$) and Boolean algebra with ``exclusive or'' as addition and ``and''108as multiplication.109\paragraph*{Boolean functions and Reed-Muller codes.}110Any Boolean function $f : \F_2^n \To \F_2$ can be represented as a reduced polynomial in at most $n$ variables over $F_2$111\cite{Mul54, Rot76} \cite[Ch. III, Section 2]{Dil74}.112This is called the \Emph{algebraic normal form} of $f$ \cite[Chapter 5]{Rue1986}.113114\begin{example}115In Sage, using \verb!sage.crypto.boolean_function!, define:116\begin{sageblock}117from sage.crypto.boolean_function import BooleanFunction118f = BooleanFunction([1,1,1,0])119a = f.algebraic_normal_form()120\end{sageblock}121The algebraic normal form of the Boolean function \verb!f! on $\F_2^2$122with variables $x_0$ and $x_1$ and truth table $\sage{[1,1,1,0]}$ (in lexicographic order)123is then \verb!a = !$\sage{a}$124\cite[Boolean Functions]{SageMath8418}.125\end{example}126127\begin{Definition}128\label{def-Reed-Muller-codes}129\cite{Mul54} \cite[Ch. 13, Section 3]{MacS77} \cite[10.5.2]{Sti07combinatorial}130The \Emph{Reed-Muller code} $RM(r, n)$ consists of those Boolean functions $f : \F_2^n \To \F_2$131whose algebraic normal form has degree $r$.132\end{Definition}133\begin{remark}134Some texts use the notation $\mathcal{R}(r,n)$ or $RM(r,2^n)$ for $RM(r,n)$.135136Each Reed-Muller code $RM(r,n)$ is a linear subspace of the vector space of Boolean functions $f : \F_2^n \To \F_2$.137138The Reed-Muller code $RM(1,n)$ consists of the $2^{n+1}$ affine functions $f(x) = \langle c, x \rangle + \delta$139for $c \in \F_2^n,\ \delta \in \F_2$ \cite[Ch 14, Section 3]{MacS77} \cite[10.5.2]{Sti07combinatorial}.140\end{remark}141142% \begin{example}143% Using \verb!sage.coding.reed_muller_code! in Sage, list the Reed-Muller code $RM(1,2)$:144% \begin{sageblock}145% from sage.coding.reed_muller_code import ReedMullerCode146% C = ReedMullerCode(GF(2),1,2)147% L = C.list()148% \end{sageblock}149% This sets the variable \verb!L! to be the list of 8 vectors150% \begin{align*}151% [ &\sage{L[0]},\sage{L[1]},\sage{L[2]},\sage{L[3]},152% \\153% &\sage{L[4]},\sage{L[5]},\sage{L[6]},\sage{L[7]}\,]154% \end{align*}155% \cite[Reed-Muller Code]{SageMath8418}.156% \end{example}157158\paragraph*{Bent Boolean functions.}159Bent Boolean functions can be defined in a number of equivalent ways.160The definition used here involves the Walsh Hadamard Transform.161\begin{Definition}162\label{def-Walsh-Hadamard-transform}163\cite[Ch. III, Section 2]{Dil74} \cite[Ch. 2, Section 3]{MacS77}164The Walsh Hada\-mard transform of165a Boolean function $f : \F_2^n \To \F_2$ is166\begin{align*}167W_f(x)168&:=169\sum_{y \in \F_2^n} (-1)^{f(y) + \langle x, y \rangle}170\intertext{where}171\langle x , y \rangle172&:= \sum_{i=0}^{n-1} x_i y_i.173\end{align*}174\end{Definition}175\begin{example}176Using \verb!sage.crypto.boolean_function! in Sage, define:177\begin{sageblock}178from sage.crypto.boolean_function import BooleanFunction179f = BooleanFunction([1,1,1,0])180w = f.walsh_hadamard_transform()181\end{sageblock}182\begin{sagesilent}183from sage.misc.banner import require_version184\end{sagesilent}185186The Walsh Hada\-mard transform of the Boolean function \verb!f! on $\F_2^2$ is then187\verb!w =! $\sage{w if require_version(8,2,0) else tuple(-v for v in w)}$188as a truth table in lexicographic order189\cite[Boolean Functions]{SageMath8418}.190\end{example}191\begin{remarks}~192193\begin{enumerate}194\item195In versions of Sage before 8.2 the \verb!walsh_hadamard_transform! method has an incorrect sign196[Sage trac ticket \#23931].197\item198For Boolean functions $f : \F_{2^n} \To \F_2$,199where $\F_{2^n}$ is the finite field on $2^n$ elements,200the \emph{trace form} \cite[3.1]{Jac64}201202is used to define the Walsh Hada\-mard transform.203\end{enumerate}204\end{remarks}205206\begin{Definition}207\label{def-Bent-function}208A Boolean function $f : \F_2^{2m} \To \F_2$ is \Emph{bent}209if and only if its Walsh Hada\-mard transform has constant absolute value $2^{m}$ \cite[p. 74]{Dil74}210\cite[p. 300]{Rot76}.211\end{Definition}212\begin{example}213The Boolean function \verb!f! in the previous example is bent, since its214Walsh Hada\-mard transform has constant absolute value $\sage{abs(w[0])}$.215\end{example}216217218The remainder of this paper refers to bent Boolean functions simply as bent functions.219\begin{remarks}~220221\begin{enumerate}222\item223Bent functions can also be characterized as those Boolean functions whose Hamming distance224from any affine Boolean function is the maximum possible \cite[Ch. 14 Theorem 6]{MacS77} \cite[Theorem 3.3]{MeiS90}.225\item226The property of being a bent function is invariant with respect to the non-degenerate symmetric bilinear form used to define227the Walsh Hada\-mard transform.228That is, for a non-degenerate symmetric bilinear form $\langle x, S y \rangle$, where $S$ is a non-degenerate symmetric matrix in $\F_2^{n \times n}$,229and a Boolean function $f : \F_{2^n} \To \F_2$, the Walsh Hada\-mard transform230\begin{align*}231W[S]_f(x)232&:=233\sum_{y \in \F_2^n} (-1)^{f(y) + \langle x, S y \rangle}234\end{align*}235has constant absolute value $2^{m}$ if and only if $f$ is bent as per Definition \ref{def-Bent-function} \cite{DinMTX2018cyclic}.236In this sense, given an isomorphism between $\F_2^{2m}$ and $\F_{2^{2m}}$ as vector spaces over $\F_2$,237the definition of a bent function on $\F_2^{2m}$ is equivalent to the definition on $\F_{2^{2m}}$.238\item239The fact that any symmetric matrix $S$ on $\F_2^n$ can be factorized $S=L^T L$, with the shape of $L$ depending on the rank of $S$ \cite{Lem1975matrix}240leads to the existence of a basis of $\F_{2^n}$ for which the trace form is diagonal.241This yields an explicit isomorphism between $\F_2^{2m}$ and $\F_{2^{2m}}$ as vector spaces over $\F_2$ for which the two equivalent definitions242of bent function coincide \cite[p. 860]{OlSW1975}.243\end{enumerate}244%245\end{remarks}246247The characterization of bent functions given by Definition~\ref{def-Bent-function} immediately248implies the existence of dual functions:249\begin{Definition}250\label{def-dual-Bent-function}251For a bent function $f : \F_2^{2m} \To \F_2$, the function $\dual{f}$, defined by252\begin{align*}253(-1)^{\dual{f}(x)} &:= 2^{-m} W_f(x)254\end{align*}255is called the \Emph{dual} of $f$ \cite{CarDPS10self}.256\end{Definition}257258\begin{Remark}259The function $\dual{f}$ is also a bent function on $\F_2^{2m}$ \cite[p. 427]{MacS77} \cite[p. 301]{Rot76}.260\end{Remark}261262\subsection{Weights and weight classes}263\begin{Definition}264\label{def-weight}265The \Emph{Hamming weight} of a Boolean function is the cardinality of its \Emph{support} \cite[p. 8]{MacS77}.266For $f$ on $\F_2^n$267\begin{align*}268\support{f} &:= \{x \in \F_2^n \mid f(x)=1 \}, \quad \weight{f} := \abs{ \support{f} }.269\end{align*}270\end{Definition}271272The remainder of this paper refers to Hamming weights simply as weights.273274Since a bent function of a given dimension can have only one of two weights,275the weights can be used to define equivalence classes of bent functions % and their Cayley graphs,276here called \emph{weight classes}.277\begin{Definition}278\label{def-weight-class}279A bent function $f$ on $\F_2^{2m}$ has weight \cite[Theorem 6.2.10]{Dil74}280\begin{align*}281\weight{f} &= 2^{2 m - 1} - 2^{m-1} \quad (\text{weight class number~} \weightclass{f}=0),282\text{~or}283\\284\weight{f} &= 2^{2 m - 1} + 2^{m-1} \quad (\text{weight class number~} \weightclass{f}=1).285\end{align*}286% If $f(0)=0$ then $\weightclass{\Cay{f}} := \weightclass{f}$.287\end{Definition}288% \end{frame}289290\paragraph*{Weight classes and dual bent functions.}291%\subsection{Weight classes and dual bent functions.}292293We now note a connection between weight classes and dual bent functions that294makes it a little easier to reason about dual bent functions.295The following lemma expresses the dual bent function in terms of weight classes.296(See also MacWilliams and Sloane \cite[p. 414]{MacS77}.)297\begin{Lemma}298\label{lm-notes-9b}299For a bent function $f : \F_2^{2m} \To \F_2$, and $x \in \F_2^{2m}$,300\begin{align*}301\dual{f}(x)302&=303\weightclass{y \mapsto f(y) + \langle x, y \rangle}.304\end{align*}305306\end{Lemma}307308The proof of Lemma~\ref{lm-notes-9b} relies on the following lemma about weight classes.309\begin{Lemma}310\label{lm-notes-9a}311For a bent function $f : \F_2^{2m} \To \F_2$,312\begin{align*}313\weightclass{f}314&=3152^{-m} \weight{f} - 2^{m-1} + 2^{-1},316\intertext{so that}317\weight{f}318&=3192^{m} \weightclass{f} + 2^{2m-1} - 2^{m-1}.320\end{align*}321322\end{Lemma}323324\begin{proof}325If $\weight{f} = 2^{2 m - 1} - 2^{m-1}$ then326\begin{align*}3272^{-m} \weight{f} - 2^{m-1} + 2^{-1}328&=3292^{-m} (2^{2 m - 1} - 2^{m-1}) - 2^{m-1} + 2^{-1} = 0.330\end{align*}331If $\weight{f} = 2^{2 m - 1} + 2^{m-1}$ then332\begin{align*}3332^{-m} \weight{f} - 2^{m-1} + 2^{-1}334&=3352^{-m} (2^{2 m - 1} + 2^{m-1}) - 2^{m-1} + 2^{-1} = 1.336\end{align*}337\end{proof}338339\begin{proofof}{Lemma~\ref{lm-notes-9b}}340Let $h(y) := y \mapsto f(y) + \langle x, y \rangle.$341Then342\begin{align*}343(-1)^{\dual{f}(x)}344&=3452^{-m} \sum_{y \in \F_2^{2m}} (-1)^{f(y) + \langle x, y \rangle}346\\347&=3482^{-m} \left( \sum_{f(y) + \langle x, y \rangle = 0} 1 - \sum_{f(y) + \langle x, y \rangle = 1} 1349\right)350\\351&=3522^{-m} \left( 2^{2m} - 2 \weight{h} \right)353=3542^m - 2^{1-m} \weight{h}355\\356&=3572^m - 2^{1-m} (2^{m} \weightclass{h} + 2^{2m-1} - 2^{m-1})358\\359&=360%2^m - 2 \weightclass{h} - 2^m + 1361%=3621 - 2 \weightclass{h} = (-1)^{\weightclass{h}},363\end{align*}364where we have used Lemma~\ref{lm-notes-9a}.365\end{proofof}366367%368%~369%370%\slidecite{Dillon 1974; Rothaus 1976; Tokareva 2011}371% \end{frame}372373% \begin{frame}374\section{Bent functions and strongly regular graphs}375\label{sec-Bent-graphs}376This section defines the Cayley graph of a Boolean function,377and explores the relationships between bent functions, projective two-weight codes, and strongly regular graphs.378\subsection{The Cayley graph of a Bent function}379380The Cayley graph of a bent function $f$ with $f(0)=0$ is defined381in terms of the Cayley graph for a general Boolean function with $f$ with $f(0)=0$.382\paragraph*{The Cayley graph of a Boolean function.}383%\begin{center}384\begin{Definition}385\label{def-Cayley-graph}386For a Boolean function $f : \F_2^{2 m} \To \F_2$, with $f(0)=0$ we consider the simple undirected387\emph{Cayley graph} $\Cay{f}$ \cite[3.1]{BerC99}388where the vertex set $V(\Cay{f}) = \F_2^{2 m}$ and for $i,j \in \F_2^{2 m}$, the edge $(i,j)$ is in389the edge set $E(\Cay{f})$ if and only if $f(i+j)=1$.390\end{Definition}391Note especially that in contrast with the paper of Bernasconi and Codenotti \cite{BerC99},392this paper defines Cayley graphs only for Boolean functions $f$ with $f(0)=0$,393since the use of Definition~\ref{def-Cayley-graph} with a function $f$ for which $f(0)=1$ would394result in a graph with loops rather than a simple graph.395396%\slidecite{Bernasconi and Codenotti 1999} % BerC99397% \end{frame}398% \begin{frame}399\paragraph*{Bent functions and strongly regular graphs.}400%\begin{center}401We repeat below in Proposition~\ref{pr-Cayley-bent-strongly-regular}402the result of Bernasconi and Codenotti \cite{BerC99}403that the Cayley graph of a bent function is strongly regular.404The following definition is used to fix the notation used in this paper.405\begin{Definition}406\label{def-strongly-regular-graph}407%408A simple graph $\Gamma$ of order $v$ is \Emph{strongly regular} \cite{Bos63,BroCN89,Sei79} with409parameters410$(v,k,\lambda,\mu)$ if411\begin{itemize}412\item413each vertex has degree $k,$414\item415each adjacent pair of vertices has $\lambda$ common neighbours, and416\item417each nonadjacent pair of vertices has $\mu$ common neighbours.418\end{itemize}419%420\end{Definition}421%~422%423%\slidecite{Brouwer, Cohen and Neumaier 1989} % BroCN89424425%\end{center}426% \end{frame}427428% \begin{frame}429The following proposition summarizes some of the well-known properties of the Cayley graphs of bent functions.430\begin{Proposition}431\label{pr-Cayley-bent-strongly-regular}432The Cayley graph $\Cay{f}$ of a bent function $f$ on $\F_2^{2m}$433with $f(0)=0$ is a strongly regular graph with $\lambda = \mu$ \cite[Lemma 12]{BerC99}.434435In addition, any Boolean function $f$ on $\F_2^{2m}$ with $f(0)=0$,436whose Cayley graph $\Cay{f}$ is a strongly regular graph with $\lambda = \mu$ is a bent function437\cite[Theorem 3]{BerCV01} \cite[Theorem 3.1]{Sta07}.438439For a bent function $f$ on $\F_2^{2m}$,440the parameters of $\Cay{f}$ as a strongly regular graph441are \cite[Theorem 6.2.10]{Dil74} \cite[Theorem 3.2]{HuaY04}442\begin{align*}443(v,k,\lambda,\mu) = &(4^m, 2^{2 m - 1} - 2^{m-1}, 2^{2 m - 2} - 2^{m-1}, 2^{2 m - 2} - 2^{m-1})444\\445\text{or} \quad &(4^m, 2^{2 m - 1} + 2^{m-1}, 2^{2 m - 2} + 2^{m-1}, 2^{2 m - 2} + 2^{m-1}).446\end{align*}447\end{Proposition}448449%~450%451%\slidecite{Menon 1962; Dillon 1974; Bernasconi and Codenotti 1999}452%\end{center}453% \end{frame}454% \begin{frame}455\subsection{Bent functions, linear codes and strongly regular graphs}456Another well known way to obtain a strongly regular graph from a bent function is via a projective two-weight457code.458This is done via the following definitions.459\paragraph*{Projective two-weight binary codes.}460461\begin{Definition}462\label{def-two-weight-codes}463\cite{BouFFWW2006, CalK1986, Del72weights, Din2015, Ton96uniformly}464465A \Emph{two-weight binary code} with parameters $[n,k,d]$ is a $k$ dimensional subspace of $\F_2^n$466with467minimum Hamming distance $d$, such that the set of Hamming weights of the non-zero vectors has size4682.469470Bouyukliev, Fack, Willems and Winne \cite[p. 60]{BouFFWW2006} define projective codes as follows.471``A \Emph{generator matrix} $G$ of a linear code $[n, k]$ code $C$ is any matrix472of rank $k$ (over $\F_2$) with rows from $C.$ \ldots473A linear $[n, k]$ code is called \Emph{projective} if no two columns of a generator matrix474$G$ are linearly dependent, i.e., if the columns of $G$ are pairwise different points in a475projective $(k-1)$-dimensional space.''476477A \Emph{projective two-weight binary code} with parameters $[n, k, d]$ is thus a478two-weight binary code with these parameters which is also projective as an $[n, k]$ linear code.479%480%~481%482%\slidecite{Bouyukliev, Fack, Willems and Winne 2006} % BouFFWW2006483%484\end{Definition}485486\begin{Remark}487In the case of $\F_2$, no two columns of the generator matrix of a projective code are equal.488\end{Remark}489490491\paragraph*{From bent function to strongly regular graph via a projective two-weight code.}492There is a standard construction for obtaining a projective two-weight code from a bent function in such a way that493the code can be used to define a strongly regular graph.494This method is equivalent to the following definition.495\begin{Definition}496\label{def-boolean-linear-code}497\cite{Din2015}.498%\smallcite{Ding 2015, Corollary 10}499See also \cite[Definition 2A]{CalK1986}500501Let $v=2^{2m}$, and let $X$ be the matrix in $\F_2^{2m \times v}$502whose columns are the $v$ elements of $\F_2^{2m}$ in lexicographic order.503For a Boolean function $f : \F_2^{2m} \To \F_2$,504let $n=\weight{f}$, and let $Y$ be the matrix in $\F_2^{2m \times n}$505whose columns are the $n$ elements of $\support{f}$ in lexicographic order.506That is,507\begin{align*}508Y_{i,j} &= (y_j)_i, \quad \text{for~} i \in \{0,\ldots,v-1\},\ j \in \{0,\ldots,n-1\},509\end{align*}510with $y_j \in \F_2^{2m}$ being an element of $\support{f}$.511512The linear code $C(f) \subset \F_2^n$ is then given by the span of the $2 m$ rows of $Y$ within $\F_2^n$,513considered as row vectors.514\end{Definition}515% define the linear code $C(f)$ by the generator matrix516% \begin{align*}517% M_C(f)_{x,y} &\in \F_2^{2^{2m} \times \weight{f}},518% \\519% M_C(f)_{x,y} &:= \langle x, \support{f}(y) \rangle,520% \end{align*}521% with $x$ in lexicographic order of $\F_2^{2m}$522% and $\support{f}(y)$ in lexicographic order of $\support{f}$.523%524% The $4^m$ words of the code $C(f)$ are the rows of the generator matrix $M_C(f)$.525526\begin{example}527In Sage, using \verb!sage.crypto.boolean_function! and528\newline529\verb!boolean_cayley_graphs.boolean_linear_code!, define:530\begin{sageblock}531from sage.crypto.boolean_function import BooleanFunction532from boolean_cayley_graphs.boolean_linear_code import (533boolean_linear_code)534f = BooleanFunction([0,1,0,0,0,1,0,0,1,0,0,0,0,0,1,1])535dim = f.nvariables()536Cf = boolean_linear_code(dim, f)537\end{sageblock}538The linear code $Cf := C(f)$ of the Boolean function \verb!f! on $\F_2^4$ then has the following539generator matrix in echelon form:540\begin{align*}541\sage{Cf.generator_matrix().echelon_form()}.542\end{align*}543\cite{Leo16GitHub} \cite[Boolean Functions]{SageMath8418}.544\end{example}545546\begin{remarks}~547548\begin{enumerate}549\item550The linear span of the rows of $Y$ is given by the set of rows of the matrix551%552\begin{align*}553M &:= X^T Y, \quad \text{where}554\\555X &\in \F_2^{2m \times v}556\end{align*}557is the matrix whose columns are all $v$ elements of $\F_2^{2m}$ in lexicographic order.558Thus $M \in \F_2^{v \times n}$.559\item560Definition \ref{def-boolean-linear-code} is not identical to the generic construction described by561Ding \cite[Section III]{Din2015} since that definition is for562subsets $D \subset \F_v$, and uses the trace form, but for the case where $D$ is the support563of a Boolean function $f \colon \F_v \To F_2$, the two definitions are equivalent.564565If the bilinear form $\langle x, y \rangle = x^T y$ is replaced in566the previous remark by a non-degenerate symmetric bilinear form $\langle x, S y \rangle = x^T S y$,567where $S$ is a non-degenerate symmetric matrix in $\F_2^{2m \times 2m}$,568then this would produce the same linear code, since the matrix $S$, being invertible,569produces a bijection on $\F_2^{2m}$ and therefore the matrix $X^T S$ differs from $X^T$ only by a permutation of rows.570571Thus by similar arguments to those in Remarks 2 and 3 following Definition \ref{def-Bent-function},572the generic construction described by Ding \cite[Section III]{Din2015} is in this case equivalent573to Definition \ref{def-boolean-linear-code}.574\end{enumerate}575\end{remarks}576577When $f$ is a bent function, the linear code $C(f)$578described by Definition \ref{def-boolean-linear-code} has the following properties.579%\slidecite{Ding 2015, Corollary 10}580%581% \end{frame}582% \begin{frame}583%\frametitle{From bent function to linear code (2)}584\begin{Proposition}585\label{pr-bent-two-weight-code}586\cite[Corollary 10]{Din2015}587%\smallcite{Ding 2015, Corollary 10}588589For a bent function $f : \F_2^{2m} \To \F_2$, the linear code $C(f)$590is a projective two-weight binary code, with591\begin{align*}592\begin{cases}593n = \weight{f} = 2^{2m-2} - 2^{m-2}, \quad k = 2m & \text{if~} \weightclass{f}=0.594\\595n = \weight{f} = 2^{2m-2} + 2^{m-2}, \quad k = 2m & \text{if~} \weightclass{f}=1.596\end{cases}597\end{align*}598%599The possible weights of non-zero code words are:600\begin{align*}601\begin{cases}6022^{2m-2}, 2^{2m-2} - 2^{m-1} & \text{if~} \weightclass{f}=0.603\\6042^{2m-2}, 2^{2m-2} + 2^{m-1} & \text{if~} \weightclass{f}=1.605\end{cases}606\end{align*}607%608\end{Proposition}609610\begin{proof}611We examine $W_f$, the Walsh Hadamard transform of $f$.612\begin{align*}613W_f(x)614&=615\sum_{y \in \F_2^{2 m}} (-1)^{\langle x, y \rangle + f(y)}616=617\sum_{y \in \F_2^{2 m}} (-1)^{\langle x, y \rangle}618- 2\sum_{f(y)=1} (-1)^{\langle x, y \rangle}.619\end{align*}620But621\begin{align*}622\sum_{y \in \F_2^{2 m}} (-1)^{\langle x, y \rangle}623&=624\begin{cases}6254^m &(x=0)626\\6270 & \text{otherwise},628\end{cases}629\end{align*}630as per the Sylvester Hadamard matrices.631632So, for $x \neq 0$,633\begin{align*}634W_f(x)635&=636- 2\sum_{f(y)=1} (-1)^{\langle x, y \rangle},637\intertext{so}638\sum_{f(y)=1} (-1)^{\langle x, y \rangle}639&=640\weight{f} - 2 \sum_{\substack{f(y)=1 \\ \langle x, y \rangle =1}} 1641=642- W_f(x)/2.643\intertext{But}644\sum_{\substack{f(y)=1 \\ \langle x, y \rangle =1}} 1645&=646\weight{x^T Y},647\end{align*}648the weight of the code word of $C(f)$ corresponding to $x$.649So650\begin{align*}651\weight{f} - 2 \weight{x^T Y}652&=653- W_f(x)/2,654\intertext{and therefore}655\weight{x^T Y}656&=657\weight{f}/2 + W_f(x)/4.658\end{align*}659We now examine the two possible weight class numbers of $f$.660661If $\weightclass{f} = 0$ then $\weight{f} = 2^{2m-1}-2^{m-1}$.662For $x \neq 0$ there are two cases, depending on $\dual{f}(x)$:663664If $\dual{f}(x) = 0$ then $W_f(x) = 2^m$, so665\begin{align*}666\weight{x^T Y}667&=6682^{2m-2} - 2^{m-2} + 2^{m-2}669=6702^{2m-2}.671\end{align*}672673If $\dual{f}(x) = 1$ then $W_f(x) = -2^m$, so674\begin{align*}675\weight{x^T Y}676&=6772^{2m-2} - 2^{m-2} - 2^{m-2}678=6792^{2m-2} - 2^{m-1}.680\end{align*}681682Similarly, if $\weightclass{f} = 1$ then $\weight{f} = 2^{2m-1}+2^{m-1}$,683and so for $x \neq 0$684\begin{align*}685\weight{x^T Y}686&=687\begin{cases}6882^{2m-2} + 2^{m-1} & (\dual{f}(x)=0)689\\6902^{2m-2} & (\dual{f}(x)=1).691\end{cases}692\end{align*}693As a consequence, all $v = 2^{2m}$ rows of $M = X^T Y$ are distinct,694since the rows of $M$ consist of every linear combination of the rows of $Y$,695every linear combination of the rows of $Y$ is given by some $x^T Y$ where $x$ is a column of $X$,696and for every non-zero $x$, this linear combination has positive weight and is therefore non-zero.697Thus the dimension of $C(f)$ as a subspace of $F_2^n$ is $k=2m$.698\end{proof}699700%\slidecite{Ding 2015, Corollary 10}701%702% \end{frame}703% \begin{frame}704\paragraph*{From linear code to strongly regular graph.}705This paper uses the following non-standard definition to obtain a strongly regular graph from a706projective two-weight code.707\begin{Definition}708\label{def-R-f}709Given $f : \F_2^{2m} \To \F_2$, and linear code $C(f)$ defined as per Definition~\ref{def-boolean-linear-code},710define the graph $R(f)$ as follows.711712Vertices of $R(f)$ are code words of $C(f)$.713714For code words $v,w \in C(f)$, edge $(u,v) \in R(f)$ if and only if715\begin{align*}716\begin{cases}717\weight{u+v} = 2^{2m-2} - 2^{m-1} & (\text{if~}\weightclass{f}=0).718\\719\weight{u+v} = 2^{2m-2} + 2^{m-1} & (\text{if~}\weightclass{f}=1).720\end{cases}721\end{align*}722723\end{Definition}724\begin{remarks}~725726\begin{enumerate}727\item728The standard definition uses the lower of the two weights in both cases above.729\item730Since $C(f)$ is a projective two-weight binary code,731$R(f)$ is a strongly regular graph \cite[Theorem 2]{Del72weights} \cite[Theorem 16.22]{CamVL91}.732\end{enumerate}733\end{remarks}734735736%\slidecite{Delsarte 1972, Theorem 2}737% \end{frame}738739% \end{frame}740741742% \end{frame}743%\subsection{Bent functions, linear codes and strongly regular graphs}744% \begin{frame}745\paragraph*{The graph $R(f)$ is the Cayley graph of the extended dual.}746The strongly regular graph $R(f)$ of bent function $f$, as per747\ref{def-R-f} has the following remarkable property.748749\begin{Theorem}750For a bent function $f : \F_2^{2m} \To \F_2$, with $f(0)=0$,751\begin{align*}752R(f) \equiv \Cay{\dual{f} + \weightclass{f}}.753\end{align*}754755\end{Theorem}756% \end{frame}757\begin{proof}758As a consequence of Lemma~\ref{lm-notes-9b}, $\weightclass{f} = \dual{f}(0)$,759so if $g(x) := \dual{f}(x) + \weightclass{f}$ then $g(0)=0$ and therefore the Cayley graph of $g$760is well defined.761762From the proof of Proposition~\ref{pr-bent-two-weight-code}:763764If $\weightclass{f} = 0$ then for $x \neq 0$765\begin{align*}766\weight{x^T Y}767&=768\begin{cases}7692^{2m-2} & (\dual{f}(x)=0)770\\7712^{2m-2} - 2^{m-1} & (\dual{f}(x)=1).772\end{cases}773\end{align*}774Thus for $x \neq 0$, if $\dual{f}(x) + weightclass{f} = 1$ if and only if $\weight{x^T Y} = 2^{2m-2} - 2^{m-1}$.775776If $\weightclass{f} = 1$ then for $x \neq 0$777\begin{align*}778\weight{x^T Y}779&=780\begin{cases}7812^{2m-2} + 2^{m-1} & (\dual{f}(x)=0)782\\7832^{2m-2} & (\dual{f}(x)=1).784\end{cases}785\end{align*}786Thus for $x \neq 0$, if $\dual{f}(x) + weightclass{f} = 1$ if and only if $\weight{x^T Y} = 2^{2m-2} + 2^{m-1}$.787788Thus in both cases, $R(f)$ as per Definition~\ref{def-R-f} is isomorphic to the Cayley graph $\Cay{\dual{f} + \weightclass{f}}$.789\end{proof}790791\section{Equivalence of bent functions}792\label{sec-Equivalence}793The following concepts of equivalence of Boolean functions are used in this paper,794usually in the case where the Boolean functions are bent.795796% \begin{frame}797\paragraph*{Extended affine equivalence.}798799\begin{Definition}800For Boolean functions $f,g : \F_2^n \To \F_2$,801$f$ is \Emph{extended affine equivalent} to $g$ \cite[Section 1.4]{Tok15bent} if and only if802\begin{align*}803g(x) &= f(A x + b) + \langle c, x \rangle + \delta804\end{align*}805for some $A \in GL(n,2)$, $b, c \in \F_2^n$, $\delta \in \F_2$.806\end{Definition}807808The Boolean function $f$ is extended affine equivalent to $g$809if and only if $f$ and $g$ are in the same orbit810of the action of the \Emph{extended general affine group} $EGA(n, 2)$ on $\F_2^{\F_2^n}$, defined as follows.811812\begin{Definition}813\begin{align*}814&EGA(n, 2) := \{ (A,b,c,\delta) \mid A \in GL(n,2),\ b, c \in \F_2^n,\ \delta \in \F_2 \}815\intertext{with}816&(A,b,c,\delta)(A',b',c',\delta') := (A A', A b' + b, A'^T c + c', \langle c, b' \rangle + \delta + \delta'),817\end{align*}818with action819\begin{align*}820(A,b,c,\delta)f(x) &:= f(A x + b) + \langle c, x \rangle + \delta,821\\822\left( (A,b,c,\delta)(A',b',c',\delta') f \right)823& := (A',b',c',\delta') \circ (A,b,c,\delta) f824\\825& = (A',b',c',\delta') \left( (A,b,c,\delta) f \right).826\end{align*}827\cite[Section 2]{Mai91}828\end{Definition}829830\begin{Proposition}831\label{prop-EA-class-properties}832The extended affine (EA) equivalence classes of the Boolean functions $\F_2^n \To \F_2$,833that is, the orbits of these functions under $EGA(n, 2)$,834have the following well known and easily verified properties.835\begin{enumerate}836\item837For a given $f \colon \F_2^n \To \F_2$,838the $2^{n+1}$ functions $x \mapsto f(x) + \langle c, x \rangle + \delta$ are all distinct.839Thus the EA equivalence class of $f$ consists of some number of complete cosets of the Reed-Muller code $RM(1,n)$840described in Section \ref{sec-Boolean-functions}.841\item842Each general affine transformation $(A,b)f(x) := f(A x + b)$ preserves cosets of the Reed-Muller code $RM(1,n)$843in the sense that $(A,b)$ maps $f + RM(1,n)$ to $g + RM(1,n)$ where $g(x) = f(A x + b)$.844\end{enumerate}845%846\end{Proposition}847See also MacWilliams and Sloane \cite[Ch. 13]{MacS77}, and Maiorana \cite{Mai91}.848% \begin{frame}849\paragraph*{General linear equivalence.}850\begin{Definition}851For Boolean functions $f,g : \F_2^n \To \F_2$,852$f$ is \Emph{general linear equivalent} to $g$ if and only if853\begin{align*}854g(x) &= f(A x)855\end{align*}856for some $A \in GL(n,2)$.857\end{Definition}858859Thus $f$ is general linear equivalent to $g$860if and only if $f$ and $g$ are in the same orbit861of the action of the general linear group $GL(n, 2)$ on $\F_2^{\F_2^n}$, defined as follows.862863\begin{Definition}864\begin{align*}865A f(x) &:= f(A x),866\\867(A A') f & := A' \circ A f = A' (A f).868\end{align*}869\end{Definition}870Some references for the study of the general linear equivalence of Boolean functions871include Harrison \cite{Har64}, Comerford \cite{Com80}, and Maiorana \cite[Section 2]{Mai91}.872% \begin{frame}873\paragraph*{Extended translation equivalence.}874875\begin{Definition}876For Boolean functions $f,g : \F_2^n \To \F_2$,877$f$ is \Emph{extended translation equivalent} to $g$ if and only if878\begin{align*}879g(x) &= f(x + b) + \langle c, x \rangle + \delta880\end{align*}881for $b, c \in \F_2^n$, $\delta \in \F_2$.882\end{Definition}883% \end{frame}884885Thus $f$ is extended translation equivalent to $g$886if and only if $f$ and $g$ are in the same orbit887of the action of the \Emph{extended translation group} $ET(n, 2)$ on $\F_2^{\F_2^n}$, defined as follows.888889\begin{Definition}890\begin{align*}891&ET(n, 2) := \{ (b,c,\delta) \mid \ b, c \in \F_2^n,\ \delta \in \F_2 \}892\intertext{with}893&(b,c,\delta)(b',c',\delta') := (b' + b, c + c', \langle c, b' \rangle + \delta + \delta'),894\end{align*}895with action896\begin{align*}897(b,c,\delta)f(x) &:= f(x + b) + \langle c, x \rangle + \delta,898\\899\left( (b,c,\delta)(b',c',\delta') \right) f900& := (b',c',\delta') \circ (b,c,\delta) f901\\902& = (b',c',\delta') \left( (b,c,\delta) f \right).903\end{align*}904\end{Definition}905906%\slidecite{Tokareva 2014}907% \end{frame}908% \begin{frame}909\paragraph*{Cayley equivalence.}910\begin{Definition}911%912For Boolean functions $f, g : \F_2^n \To \F_2$, with $f(0)=g(0)=0$,913we call $f$ and $g$ \Emph{Cayley equivalent},914and write $f \equiv g$,915if and only if the graphs $\Cay{f}$ and $\Cay{g}$ are isomorphic.916917Equivalently, $f \equiv g$ if and only if918there exists a bijection $\pi : \F_2^n \To \F_2^n$ such that919\begin{align*}920g(x+y) &= f \big(\pi(x)+\pi(y)\big) \quad \text{for all~} x,y \in \F_2^n.921\end{align*}922\end{Definition}923% \end{frame}924Remark: Note that the bijection $\pi$ is not necessarily linear on $\F_2^n$.925Examples of bent functions $f$ and $g$ where $f \equiv g$ but the bijection is not linear926are given in Section~\ref{sec-Empirical}.927% \begin{frame}928\paragraph*{Extended Cayley equivalence.}929%930While Bernasconi and Codenotti \cite{BerC99} define Cayley graphs for Boolean functions with931$f(0)=1$ and allow Cayley graphs to have loops, this paper defines Cayley932graphs only for Boolean functions where $f(0)=0$.933This has the disadvantage that Cayley equivalence is an equivalence relation on half934of the Boolean functions rather than all of them.935To extend this equivalence relation to all Boolean functions,936we just declare the functions $f$ and $f+1$ to be ``extended'' Cayley equivalent,937resulting in the following definition.938\begin{Definition}939For Boolean functions $f, g : \F_2^n \To \F_2$,940if there exist $\delta, \epsilon \in \{0,1\}$ such that $f + \delta \equiv g + \epsilon$,941we call $f$ and $g$ \Emph{extended Cayley (EC) equivalent} and write $f \cong g$.942\end{Definition}943Extended Cayley equivalence is thus an equivalence relation on the set of all Boolean functions on944$\F_2^n$.945It is easy to verify that $f \cong g$ if and only if $f+f(0) \equiv g+g(0)$.946947% \end{frame}948949% \begin{frame}950\subsection{Relationships between different concepts of equivalence}951%\section{Theoretical results}952%\label{sec-Results}953% \begin{frame}954955% This section contains a number of theoretical results that serve a few purposes.956% Firstly,957As stated in the Introduction, in order to classify bent functions by their Cayley graphs,958it helps to understand the relationship between Cayley equivalence and other concepts of equivalence959of Boolean functions, especially if this helps to cut down the search space needed for the classification.960This section lists a few of these useful relationships.961962% A similar consideration applies to the duals of bent functions.963% Secondly, some empirical observations made in the classification of bent functions in small964% dimensions can be explained by theoretical results.965% Thirdly, theoretical results can improve our understanding of the relationships between966% some of the concepts introduced in the previous section, notably dual bent functions, SDP designs,967% projective two-weight codes, and strongly regular graphs.968969\paragraph*{General linear equivalence implies Cayley equivalence.}970971Firstly, general linear equivalence of Boolean functions implies Cayley equivalence.972Specifically, the following result applies.973\begin{Theorem}974\label{th-Linear-Cayley}975If $f$ is a Boolean function with $f(0)=0$ and $g(x) := f(A x)$ where $A \in GL(n,2)$,976then $f \equiv g$.977\end{Theorem}978\begin{proof}979\begin{align*}980g(x+y) &= f\big(A(x+y)\big) = f(A x + A y)\quad \text{for all~} x,y \in \F_2^n.981\end{align*}982\end{proof}983Thus, for bent functions, the following result holds.984\begin{Corollary}985\label{corr-bent-Linear-Cayley}986If $f$ is bent with $f(0)=0$ and $g(x) := f(A x)$ where $A \in GL(n,2)$,987then $g$ is bent with $g(0)=0$ and $f \equiv g$.988\end{Corollary}989Thus if $f$ is bent with $f(0)=0$, and $g$ is bent with $g(0)=0$, and $f \not\equiv g$,990then $f$ is not general linear equivalent to $g$.991This result immediately leads to another corollary.992Here, and later in this paper, we make use of the following terminology.993\begin{Definition}994A Boolean function $f \colon \F_2^n \To \F_2$ is said to be \Emph{prolific} if995there is no pair $b, c \in \F_2^n$ with $g(x) = f(x+b) + \langle c, x \rangle + f(b)$ such that $f \cong g$.996Thus the number of extended Cayley classes in the extended translation class of a prolific Boolean function997is $4^n$.998\end{Definition}9991000\begin{Corollary}1001\label{corr-no-Cayley-no-Linear}1002If $f \colon \F^{2m} \To \F_2$ is bent with $f(0)=0$, and $f$ is prolific, then there is no triple1003$A \in GL(2m,2)$, $b, c \in \F_2^{2m}$ with $f(Ax) = f(x+b) + \langle c, x \rangle + f(b)$.1004\end{Corollary}10051006% \end{frame}10071008% \begin{frame}1009\paragraph*{Extended affine, translation, and Cayley equivalence.}10101011Secondly, if $f$ is a Boolean function,1012and $h$ is a Boolean function $h$ that is extended affine equivalent to $f$,1013then a Boolean function $g$ exists that is general linear equivalent to $h$1014and extended translation equivalent to $f$:1015\begin{Theorem}1016\label{th-Affine-Translate-Linear}1017For $A \in GL(n,2)$, $b, c \in \F_2^n$, $\delta \in \F_2$,1018$f : \F_2^n \To \F_2$,1019the function1020\begin{align*}1021h(x) &:= f(A x + b) + \langle c, x \rangle + \delta1022\intertext{can be expressed as $h(x) = g(A x)$ where}1023g(x) &:= f(x+b) + \langle (A^{-1})^T c, x \rangle + \delta.1024\end{align*}1025\end{Theorem}1026% \end{frame}1027\begin{proof}1028Let $y:= A x$. Then1029\begin{align*}1030g(A x) = g(y)1031&= f(y+b) + \langle (A^{-1})^T c, y \rangle + \delta1032\\1033&= f(y+b) + \langle c, A^{-1} y \rangle + \delta1034\\1035&= f(A x + b) + \langle c, x \rangle + \delta = h(x).1036\end{align*}1037\end{proof}10381039\begin{Corollary}1040\label{corr-Affine-Translate-Cayley}1041If $f$ is a bent Boolean function,1042and a bent function $h$ is extended affine equivalent to $f$,1043then a bent function $g$ can be found that is extended Cayley equivalent to $h$1044and extended translation equivalent to $f$.1045\end{Corollary}1046% \end{frame}1047\begin{proof}1048Let $f$, $g$, and $h$ be as per Theorem \ref{th-Affine-Translate-Linear}.1049If $f$ is bent, then so are $g$ and $h$.1050Since, by Theorem \ref{th-Affine-Translate-Linear}, $g$ is general linear equivalent to $h$,1051by Theorem \ref{th-Linear-Cayley}, $g$ is extended Cayley equivalent to $h$.1052\end{proof}10531054As a consequence, to determine which strongly regular graphs occur, corresponding to each1055extended Cayley equivalence classes within the extended affine1056equivalence class of a bent function $f : \F_2^{2m} \To \F_2$ with $f(0)=0$,1057we need only examine the extended translation equivalent functions of the form1058\begin{align*}1059f(x+b) + \langle c, x \rangle + f(b),1060\end{align*}1061for each $b, c \in \F_2^{2m}$.1062This cuts down the required search space considerably.10631064% \end{frame}10651066\paragraph*{Quadratic bent functions have only two extended Cayley classes.}1067Finally, in the case of quadratic bent functions, there is a complete classification in terms of1068weight classes.1069\begin{Theorem}1070\label{th-Quadratic-Classes}1071For each $m>0$, the extended affine equivalence class of quadratic bent functions1072$q : \F_2^{2m} \To \F_2$ contains exactly two extended Cayley equivalence classes,1073corresponding to the two possible weight classes of1074$x \mapsto q(x+b) + \langle c, x \rangle + q(b)$.1075\end{Theorem}10761077The proof of this theorem is given in Appendix~\ref{app-proof-of}.10781079\subsection{Relationships between duality of bent functions and different concepts of equivalence}10801081The following propositions are based on well known results,1082but are useful in understanding the relationship1083between the duality of bent functions and various concepts of equivalence.1084% \begin{frame}1085%\subsection{Dual functions}10861087Firstly, general linear equivalence of bent functions $f$ and $g$1088implies general linear equivalence of their duals, $\dual{f}$ and $\dual{g}$,1089which implies Cayley equivalence of $\dual{f}$ and $\dual{g}$.1090\begin{Proposition}1091\label{prop-dual-linear-equivalence}1092\cite[Remark 6.2.7]{Dil74}10931094For a bent function $f : \F_2^{2m} \To \F_2$, and $A \in GL(2 m, 2)$, if1095\begin{align*}1096g(x) &:= f(A x)1097\intertext{then}1098\dual{g}(x) &= \dual{f}\big((A^T)^{-1} x \big),1099\end{align*}1100and therefore by Theorem \ref{th-Linear-Cayley}, $\dual{g} \equiv \dual{f}$.11011102If, in addition, $f=\dual{f}$ then $\dual{g} \equiv g$.1103\end{Proposition}11041105Remark: Functions of the form1106\begin{align*}1107f(x) := \sum_{k=0}^{m-1} x_{2k} x_{2k+1}1108\end{align*}1109are self dual bent functions, $f=\dual{f}$ \cite[Remark 6.3.2]{Dil74}.1110There are many other self dual bent functions \cite{CarDPS10self,FeuSSW2013}.11111112Secondly, the following proposition displays a relationship between the extended translation1113class of a bent function $f$, and that of its dual $\dual{f}$.1114\begin{Proposition}1115\label{prop-dual-affine-equivalence}1116\cite[Remark 6.2.7]{Dil74} \cite[Proposition 8.7]{Car10boolean}.1117%\smallcite{Carlet 2007, Proposition 4}1118%1119%~11201121For a bent function $f$ on $\F_2^{2m}$, and $b,c \in \F_2^{2m}$,1122if1123\begin{align*}1124g(x) &:= f(x+b) + \langle c, x \rangle1125\intertext{then}1126\dual{g}(x) &= \dual{f}(x+c) + \langle b, x \rangle + \langle b, c \rangle.1127\end{align*}1128\end{Proposition}11291130This result has an implication for the relationship between the set of bent functions within1131an extended translation (ET) equivalence class, and the set of their duals.1132Recall that a bent function is not necessarily extended affine (EA) equivalent to its dual1133\cite{LanLM08Kasami}.1134The following ``all or nothing'' property holds within an extended translation equivalence class of bent functions.1135\begin{Corollary}1136\label{cor-dual-ET-EC}1137For bent functions $f, g$ on $\F_2^{2m}$,1138if $f$ is EA equivalent to $\dual{f}$ and $g$ is ET equivalent to $f$,1139then $\dual{g}$ is EA equivalent to $g$.1140Thus, by Corollary~\ref{corr-Affine-Translate-Cayley},1141the set of isomorphism classes of Cayley graphs of the \Emph{duals} of the bent functions in1142the ET class of $f$ equals the set of isomorphism classes of Cayley graphs of1143the bent functions themselves.11441145Conversely, for a bent function $f$ on $\F_2^{2m}$,1146if there is any bent function $g$ that is ET equivalent to $f$,1147such that $\dual{g}$ is not EA equivalent to $g$, then no bent function in the ET class is EA1148equivalent to its dual, including $f$ itself.1149\end{Corollary}11501151% \begin{frame}11521153% \begin{frame}1154\section{Bent functions and block designs}1155\label{sec-Bent-designs}1156% \begin{frame}1157%\subsection{Weight classes, dual functions, and SDP designs}11581159This section examines the relationships between bent functions and symmetric block designs.11601161% \end{frame}11621163% \end{frame}1164% \begin{frame}1165%\subsection{Dual functions}11661167%~1168%1169%\slidecite{Carlet, Danielson, Parker and Sol\'e 2008; Feulner, Sok, Sol\'e and Wassermann 2011} %1170%FeuSSW20131171% \end{frame}1172\subsection{The two block designs of a bent function}11731174The first block design of a bent function $f$ on $\F_2^{2m}$ is obtained by interpreting1175the adjacency matrix of $\Cay{f}$ as the incidence matrix of a block design.1176In this case we do not need $f(0)=0$ \cite[p. 160]{DilS87block}.11771178The second block design of a bent function $f$ involves the1179\Emph{symmetric difference property}, which was first investigated by Kantor1180\cite[Section 5]{Kan75symplectic}.1181\begin{Definition}1182\label{def-Symmetric-difference-property}1183\cite[p. 49]{Kan75symplectic}.11841185A symmetric block design $\mathcal{D}$ has the symmetric difference property (SDP)1186if, for any three blocks, $B, C, D$ of $\mathcal{D},$ the symmetric difference1187$B \bigtriangleup C \bigtriangleup D$ is either a block or the complement of a block.1188\end{Definition}11891190This second block design is defined as follows.1191\begin{Definition}1192\label{def-SDP-design}1193For a bent function $f$ on $\F_2^{2m}$, define the matrix $M_D(f) \in \F_2^{2^{2m} \times 2^{2m}}$ where1194\begin{align}1195M_D(f)_{c,x} &:= f(x) + \langle c, x \rangle + \dual{f}(c),1196\label{D-f-def}1197\end{align}1198and use it as the incidence matrix of a symmetric block design, which1199we call it the \emph{SDP design} of $f$.1200\end{Definition}12011202Kantor describes the special case where $f$ is quadratic1203\cite[Section 5]{Kan75symplectic},1204and Dillon and Schatz \cite{DilS87block} describe the general case.1205See also Cameron and van Lint \cite[pp. 77-78 and Ex. 13, p. 152]{CamVL91}.12061207The following properties of SDP designs of bent functions are well known.1208\begin{Proposition}1209\label{prop-SDP-design}1210\cite[p. 160]{DilS87block} \cite[Theorem 3.29]{Neu06bent}12111212For any bent function $f$ on $\F_2^{2m}$, the SDP design of $f$ has the symmetric difference property.1213\end{Proposition}12141215\begin{Proposition}1216\label{prop-SDP-design-affine-equivalence}1217\cite[p. 161]{DilS87block} \cite{Kan83exponential}12181219For bent functions $f, g$ on $\F_2^{2m}$,1220the two SDP designs $D(f)$ and $D(g)$ are isomorphic as symmetric block designs1221if and only if $f$ and $g$ are affine equivalent.1222\end{Proposition}12231224%1225%~1226%1227%\slidecite{Dillon and Schatz 1987; Neumann 2006}1228% \end{frame}12291230% \begin{frame}1231\paragraph*{Weight classes and the SDP design matrix.}12321233%With Lemma~\ref{lm-notes-9b} in hand, it is easy to prove Lemma~\ref{lm-SDP-design-rows} on1234%the equivalence of the definitions of the SDP design of a bent function $f$ on $Z_2^{2m}$.12351236Definition~\ref{def-SDP-design} is different from1237but equivalent to the one given by Dillon and Schatz \cite[p. 160]{DilS87block}:1238\begin{Lemma}1239\label{lm-SDP-design-rows}1240\cite[3.29]{Neu06bent}12411242For any bent function $f$ on $\F_2^{2m}$, the rows of the incidence matrix $M_D(f)$1243are given by the words of minimum weight in the code spanned by the support of $f$ and the Reed-Muller code $RM(1,2m)$.1244\end{Lemma}1245(Here we have used an ordering of the elements of $\F_2^{2m}$ to define an ordering of the columns of the incidence matrix.)12461247%The proof of Lemma~\ref{lm-SDP-design-rows} is deferred to Section~\ref{sec-Results}.12481249%\begin{proofof}{Lemma~\ref{lm-SDP-design-rows}}1250\begin{proof}1251Firstly, as mentioned in Section \ref{sec-Boolean-functions},1252the Reed-Muller code $RM(1,2m)$ consists of the words spanned by the affine functions on $Z_2^{2m}$.1253Thus, the incidence matrix $M_{RM(1,2m)}$ is defined by1254\begin{align*}1255{M_{RM(1,2m)}}_{c,x} &:= \langle c, x \rangle + d,1256\end{align*}1257where $d \in \F_2$.12581259Therefore the incidence matrix of the code spanned by the support of $f$ and $RM(1,2m)$ is defined by1260\begin{align*}1261{M_{f,RM(1,2m)}}_{c,x} &:= f(x) + \langle c, x \rangle + d.1262\end{align*}1263Finally, from Lemma~\ref{lm-notes-9b} we know that1264\begin{align*}1265\weightclass{x \mapsto f(x) + \langle c, x \rangle}1266&=1267\dual{f}(c),1268\intertext{so that}1269\weightclass{x \mapsto f(x) + \langle c, x \rangle + \dual{f}(c)}1270&=12710.1272\end{align*}1273%\end{proofof}1274\end{proof}12751276The following characterization of the SDP design of a bent function $f$ also relies on1277Lemma~\ref{lm-notes-9b} for its proof.1278We first define the matrix of weight classes corresponding to the extended translation class of $f$.1279\begin{Definition}1280\label{def-weight-class-matrix}12811282For a bent function $f : \F_2^{2m} \To \F_2$,1283define the \Emph{weight class matrix} of $f$ by1284\begin{align*}1285M_{wc}(f)_{c,b}1286&:=1287\weightclass{x \mapsto f(x+b) + \langle c, x \rangle + f(b)}1288\end{align*}1289for $b,c \in \F_2^{2m}$.1290\end{Definition}12911292\begin{Theorem}1293\label{th-Dillon-Schatz}1294The weight class matrix of $f$ as given by Definition~\ref{def-weight-class-matrix}1295equals the incidence matrix of the SDP design of $f$.1296Specifically,1297\begin{align*}1298M_{wc}(f)_{c,b}1299&=1300f(b) + \langle c, b \rangle + \dual{f}(c)1301\\1302&=1303M_D(f)_{c,b},1304\end{align*}1305where $M_D(f)$ is defined by \eqref{D-f-def}.1306\end{Theorem}13071308\begin{proof}1309Let $g(x) := f(x+b) + \langle c, x \rangle + f(b)$.1310Then by change of variable $y:=x+b$,1311\begin{align*}1312\weightclass{g}1313&=1314\weightclass{y \mapsto f(y) + \langle c, y \rangle + \langle c, b \rangle + f(b)}1315\\1316&=1317\weightclass{y \mapsto f(y) + \langle c, y \rangle} + \langle c, b \rangle + f(b)1318\\1319&=1320\dual{f}(c) + \langle c, b \rangle + f(b),1321\end{align*}1322as a consequence of Lemma~\ref{lm-notes-9b}.1323\end{proof}132413251326%\newpage1327\section{SageMath and CoCalc code}1328\label{sec-Code}1329% \begin{frame}[fragile]1330%\subsection{Public worksheet on CoCalc}1331The computational results listed in this paper were obtained by the1332use of code written in Sage \cite{JoyEtAl13Sage} \cite{SageMath7517} and Python.1333This code base is called \texttt{Boolean-Cayley-graphs} and it is available both as a GitHub1334repository \cite{Leo16GitHub} and as a public CoCalc \cite{CoCalc} folder1335\cite{Leo17CoCalc}.13361337For an introduction to other aspects of coding theory and cryptography in Sage,1338see the article by Joyner et al. \cite{JoyEtAl13Sage}.13391340\paragraph*{Description of the Sage code.}13411342This section contains a brief description of some of the code included in1343\texttt{Boolean-Cayley-graphs}.1344More detailed documentation is being developed and this is intended to be included as part of the code base.1345The code itself is subject to review and revision, and may change as a result of the advice of1346those more experienced with Sage code.1347The description in this section applies to the code base as it exists in January 2018.13481349The code base is structured as a set of Sage script files. % in the \texttt{sage-code} subdirectory.1350These in turn use Python scripts, found in a subdirectory called \texttt{Boolean\_Cayley\_graphs}.1351% of \texttt{sage-code}.13521353The Python code is used to define a number of useful Python classes.1354The key class is \texttt{BentFunctionCayleyGraphClassification}.1355This class is used to store the classification of Cayley graphs within the extended translation1356class of a given bent function $f$, as well as the classification of Cayley graphs of the duals of1357each function in the extended translation class.13581359The class therefore contains the algebraic normal form of the given bent function1360%(as attribute \texttt{algebraic\_normal\_form}),1361a list of graphs1362%(\texttt{cayley\_graph\_class\_list})1363stored as strings obtained via the \texttt{graph6\_string} \cite{McKP13nauty}1364method of the \texttt{Graph} class, and two matrices,1365%(\texttt{bent\_cayley\_graph\_index\_matrix} and \texttt{bent\_cayley\_graph\_index\_matrix})1366used to store the list indices corresponding to1367the Cayley graph for each bent function in the extended translation class, and the dual of each bent1368function, respectively.1369The class also contains the weight class matrix1370%(\texttt{weight\_class\_matrix})1371corresponding to the given bent function.13721373The class is initialized by enumerating the bent functions of the form1374$x \mapsto f(x+b) + \langle c, x \rangle + f(b)$,1375and determining the Cayley graph of each.1376For each Cayley graph, the \texttt{Graph} method \texttt{canonical\_label} is used1377to invoke the Bliss package \cite{JunK07Bliss,JunK11conflict} to calculate the canonical label1378of the graph, and then \texttt{graph6\_string} is used to obtain a string.1379Each new graph is compared for isomorphism to each of the graphs in the current list,1380by simply comparing the string against each of the existing strings.1381If the new graph is not isomorphic to any existing graph, it is added to the list.1382Each list of pairwise non-isomorphic graphs can be checked by a function called \texttt{check\_graph\_class\_list}1383which uses the Nauty package to check the non-isomorphism \cite{McKP13nauty,McKP14practical}.13841385It is the efficiency of the Bliss canonical labelling algorithm, and the speed of its implementation, that makes this approach feasible.1386Even so, for an 8 dimensional bent function, the initialization of its Cayley graph classification1387can take more than 24 hours on an Intel\textregistered Core\texttrademark i7 CPU 870 running at 2.93 GHz.1388For this reason, each computed classification is saved, and a class method (\texttt{load\_mangled})1389is provided to load existing saved classifications.13901391% \end{frame}1392\paragraph*{History of the Sage code.}13931394The Sage code originated in 2015 as a series of worksheets on Sage\-Math\-Cloud (now CoCalc).1395While these were useful for investigating extended Cayley classes for bent functions in up to 6 dimensions,1396they were too slow to use for bent functions in 8 dimensions.13971398The \texttt{Boolean-Cayley-graphs} GitHub project \cite{Leo16GitHub} and public Sage\-Math\-Cloud folder \cite{Leo17CoCalc} were begun in 20161399with the intention of refactoring the code to make it fast enough to use for bent functions in 8 dimensions1400up to degree 3.1401The use of canonical labelling made this possible.14021403Further improvements were made in 2017 to enable the classification of any bent function in 8 dimensions or less1404to be computed in a reasonable time on a commodity personal computer.1405In late 2017, code was added so that the Cayley graph classifications could be accessed via a relational database \cite{Leo18Database},1406with implementations using SQLite3 \cite{SQLite} and PostgreSQL \cite{PostgreSQL}.1407Also, parallel versions of the classification functions were written using MPI4Py,1408and used on the NCI Raijin supercomputer to complete the classifications for CAST-128 and compute1409the classifications of the $\mathcal{PS}^{(+)}$ bent functions in 8 dimensions.14101411In 2018, the code was continually improved, and computation of the1412classifications of the $\mathcal{PS}^{(+)}$ bent functions in 8 dimensions was begun.14131414% \begin{itemize}1415% \item 2015-04 to 2015-05 CoCalc: Cliques-Automorphisms project starts looking at Cayley1416% classes for bent functions of dimension up to 6, using BooleanFunction and \verb!is_isomorphic()!.1417% \item 2015-12 CoCalc: Cliques-Automorphisms project1418% worksheets produced initial results used in the ACCMCC presentation.1419% The worksheets were too slow to effectively tackle bent functions in 8 dimensions.1420% \item 2016-07 SageMath: Downloaded Sage and began refactoring worksheets into Sage code.1421% \item 2016-08 GitHub: Uploaded refactored Sage code to new Boolean-Cayley-graphs project.1422% \item 2016-08 SageMath: Began using canonical labels rather than directly testing for isomorphism1423% between Cayley graphs.1424% Canonical labelling uses the Bliss algorithm, speeding up computation in comparison to the default1425% Sage algorithm,1426% and allows comparison between graphs using equality of canonically labelled graphs rather than1427% isomorphism, also giving a speed boost.1428% This finally made it feasible to check bent functions in 8 dimensions up to degree 3.1429% \item 2016-09 SageMath: Changed many Sage code files into Python modules.1430% Introduced a BentFunction class.1431% \item 2016-11 SageMath: \verb!check_graphs_using_gap!1432% \end{itemize}1433%\newpage1434\section{Discussion}1435\label{sec-Discussion}1436% \begin{frame}1437%\subsection{Questions (1)}1438The investigation of the extended Cayley classes of bent functions is just beginning, and there are many open questions.1439This section lists some of these questions.14401441The following questions have been settled only for dimensions 2, 4 and 6.1442\begin{enumerate}1443\item1444How many extended Cayley classes are there for each dimension?1445Are there ``Exponential numbers'' of classes \cite{Kan83exponential}?1446\item1447In $n$ dimensions,1448which extended translation classes contain the maximum number, $4^n$, of different extended Cayley classes?1449\item1450Which extended Cayley classes overlap more than one extended translation class?1451\item1452Which bent functions are Cayley equivalent to their dual?1453\end{enumerate}14541455In dimension 8, what are the extended affine and extended Cayley classes of bent functions of degree 4 \cite{LanL11counting}?14561457Finally, how does the concept of extended Cayley classes of bent functions generalize to bent functions1458over number fields of prime order $p \neq 2$ \cite{CheTZ11}?14591460% \slidecite{Kantor 1983; Jungnickel and Tonchev 1991; Langevin and Leander 2008, 2011}1461% \end{frame}14621463%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%14641465\paragraph*{Acknowledgements.}1466%\small{}1467This work was begun in 2014 while the author was a Visiting Fellow at the Australian National University,1468continued while the author was a Visiting Fellow and a Casual Academic at the University of Newcastle, Australia,1469and concluded while the author was an Honorary Fellow at the University of Melbourne,1470and an employee of the Bureau of Meteorology.14711472Thanks to Christine Leopardi for her hospitality at Long Beach.1473Thanks to1474Robert Craigen,1475Joanne Hall,1476Kathy Horadam,1477David Joyner,1478Philippe Langevin,1479William Martin,1480Padraig {\'O} Cath{\'a}in,1481Judy-anne Osborn,1482Dima Pasechnik and1483William Stein1484for valuable questions, discussions and advice;1485David Joyner and Caroline Melles for suggestions for improvements based on the first draft of this paper;1486Nathan Clisby, for the opportunity to become an Honorary Fellow at the University of Melbourne;1487Kodlu, a member of MathOverflow, for asking Philippe Langevin about the affine classification1488of bent functions of degree 4 in 8 dimensions;1489Gray Chan for a citation tool for IETF RFCs;1490and finally, thanks to the authors of SageMath, Bliss, and Nauty for these valuable tools1491without which I would not have been able to conduct this research.1492%\normalsize{}1493%Thanks also to the anonymous reviewer of the previous draft of this paper.14941495\newpage14961497\appendix14981499\section{Proof of Theorem \ref{th-Quadratic-Classes}}1500\label{app-proof-of}15011502The proof of Theorem \ref{th-Quadratic-Classes} relies on a number of supporting lemmas,1503which are stated and proved here.1504\begin{Lemma}1505\label{lm-notes-5}1506Let $q(x) := x^T L x$ where $L \in \F_2^{2 m \times 2 m}$,1507\begin{align*}1508L1509&:=1510\left[1511\begin{array}{cc}15120 & I1513\\15140 & 01515\end{array}1516\right],1517\intertext{so that}1518q(x) &= \sum_{k=0}^{m-1} x_k x_{m+k}.1519\end{align*}15201521Let $f(x) := q(x+b) + \langle c,x \rangle + q(b)$.1522Then there exists $c' \in \F_2^{2m}$ such that1523\begin{align*}1524f(x)1525&=1526q(x) + \langle c',x \rangle.1527\end{align*}15281529\end{Lemma}15301531\begin{proof}1532\begin{align*}1533q(x) = x^T L x, \quad \text{so~}1534q(x+b)1535&=1536(x^T+b^T) L (x+b)1537\\1538&= q(x) + x^T L b + b^T L x + q(b)1539\\1540&= q(x) + \langle (L + L^T) b, x \rangle + q(b),1541\intertext{and therefore}1542q(x+b) + \langle c, x \rangle + q(b)1543&=1544q(x) + \langle (L+L^T) b + c, x \rangle.1545\end{align*}15461547\end{proof}15481549\begin{Lemma}1550\label{lm-notes-3}1551Let $Z \in \F_2^{2 m \times 2 m}$ be symmetric with zero diagonal.1552In other words, $Z = Z^T$, $\diag{Z} = 0$.1553Then for any $M \in \F_2^{2 m \times 2 m}$,1554\begin{align*}1555x^T (M + Z) x &= x^T M x1556\end{align*}1557for all $x \in \F_2^{2 m}$.1558\end{Lemma}15591560\begin{proof}1561Let $Z$, $x$ be as above.1562Then1563\begin{align*}1564x^T Z x1565&=1566\sum_{i=0}^{2m-1} \sum_{j=0}^{2m-1} x_i Z_{i,j} x_j1567\\1568&=1569\sum_{i=0}^{2m-1} \sum_{j<i} x_i Z_{i,j} x_j\, +1570\sum_{i=0}^{2m-1} x_i Z_{i,i} x_i\, +1571\sum_{i=0}^{2m-1} \sum_{j>i} x_i Z_{i,j} x_j1572\\1573&=1574\sum_{i=0}^{2m-1} \sum_{j<i} x_i (Z_{i,j} + Z_{j,i})1575= 0.1576\intertext{Therefore}1577x^T (M + Z) x &= x^T M x + x^T Z x = x^T M x.1578\end{align*}1579\end{proof}15801581\begin{Lemma}1582\label{lm-notes-4}1583Let $q$ be defined as per Lemma \ref{lm-notes-5}.1584Then for all $c \in Z_2^{2 m}$ with $q(c)=0$, there exists $A \in GL(2 m, 2)$ such that1585\begin{align*}1586q(A x) &= q(x) + \langle c, x \rangle.1587\end{align*}1588\end{Lemma}15891590\begin{proof}1591Let $C \in \F_2^{2 m \times 2 m}$ be such that $C_{i,j} = \delta_{i,j} c_i$, where $\delta$ is the1592\Emph{Dirac delta}: $\delta_{i,j}=1$ if $i=j$ and $0$ otherwise.1593In other words $\diag{C} = c$.1594Then1595\begin{align*}1596\langle c, x \rangle1597&=1598\sum_{i=0}^{2m-1} c_i x_i1599\\1600&=1601\sum_{i=0}^{2m-1} x_i c_i x_i1602=1603x^T C x.1604\end{align*}1605Therefore, by Lemma \ref{lm-notes-3},1606\begin{align*}1607q(x) + \langle c, x \rangle1608&=1609x^T (L + Z + C) x,1610\end{align*}1611where $Z \in \F_2^{2 m \times 2 m}$ is symmetric with zero diagonal.16121613For such $Z$, let $S := Z + C$.1614We want to find $A \in \F_2^{2 m \times 2 m}$ such that $q(A x) = q(x) + \langle c, x \rangle.$1615In other words,1616\begin{align*}1617q(A x)1618&=1619(A x)^T L (A x)1620=1621x^T A^T L A x1622=1623x^T (L + S) x.1624\end{align*}1625This will be true if $A^T L A = L + S.$16261627Let1628\begin{align*}1629A1630&:=1631\left[1632\begin{array}{cc}1633A_{0,0} & A_{0,1}1634\\1635A_{1,0} & A_{1,1}1636\end{array}1637\right],1638\quad1639S1640&:=1641\left[1642\begin{array}{cc}1643S_{0,0} & S_{0,1}1644\\1645S_{0,1}^T & S_{1,1}1646\end{array}1647\right]1648=:1649\left[1650\begin{array}{cc}1651Z_{0,0} + C_{0,0} & Z_{0,1}1652\\1653Z_{0,1}^T & Z_{1,1} + C_{1,1}1654\end{array}1655\right].1656\end{align*}1657Since1658\begin{align*}1659L A1660&=1661\left[1662\begin{array}{cc}16630 & I1664\\16650 & 01666\end{array}1667\right]1668\left[1669\begin{array}{cc}1670A_{0,0} & A_{0,1}1671\\1672A_{1,0} & A_{1,1}1673\end{array}1674\right]1675=1676\left[1677\begin{array}{cc}1678A_{1,0} & A_{1,1}1679\\16800 & 01681\end{array}1682\right],1683\end{align*}1684we require that1685\begin{align*}1686A^T L A1687&=1688\left[1689\begin{array}{cc}1690A_{0,0} & A_{1,0}1691\\1692A_{0,1} & A_{1,1}1693\end{array}1694\right]1695\left[1696\begin{array}{cc}1697A_{1,0} & A_{1,1}1698\\16990 & 01700\end{array}1701\right]1702\\1703&=1704\left[1705\begin{array}{cc}1706A_{0,0}^T A_{1,0} & A_{0,0}^T A_{1,1}1707\\1708A_{0,1}^T A_{1,0} & A_{0,1}^T A_{1,1}1709\end{array}1710\right]1711\\1712&=1713L + S1714=1715\left[1716\begin{array}{cc}1717S_{0,0} & I + S_{0,1}1718\\1719S_{0,1}^T & S_{1,1}1720\end{array}1721\right],1722\end{align*}1723and therefore1724\begin{align*}1725A_{0,0}^T A_{1,0}1726&=1727S_{0,0},1728\quad1729A_{0,0}^T A_{1,1}1730=1731I + S_{0,1},1732\\1733A_{0,1}^T A_{1,0}1734&=1735S_{0,1}^T,1736\quad1737A_{0,1}^T A_{1,1}1738=1739S_{1,1}.1740\end{align*}1741If $S_{0,1}=0$ and $A_{0,0}=I$ then1742$A_{1,0}=S_{0,0}$, $A_{1,1}=I$ and $A_{0,1}=S_{1,1}$.1743In this case, we have $A_{0,1}^T A_{1,0} = S_{0,1}^T = 0$,1744i.e. $S_{1,1} S_{0,0} = 0$, and1745\begin{align*}1746A1747&=1748\left[1749\begin{array}{cc}1750I & S_{1,1}1751\\1752S_{0,0} & I1753\end{array}1754\right],1755\intertext{so that}1756A^T L A1757&=1758\left[1759\begin{array}{cc}1760I & S_{0,0}1761\\1762S_{1,1} & I1763\end{array}1764\right]1765\left[1766\begin{array}{cc}1767S_{0,0} & I1768\\17690 & 01770\end{array}1771\right]1772\\1773&=1774\left[1775\begin{array}{cc}1776S_{0,0} & I1777\\17780 & S_{1,1}1779\end{array}1780\right]1781\\1782&=1783L + S.1784\end{align*}17851786Also1787\begin{align*}1788S1789&=1790\left[1791\begin{array}{cc}1792Z_{0,0} + C_{0,0} & 01793\\17940 & Z_{1,1} + C_{1,1}1795\end{array}1796\right].1797\end{align*}17981799Since $q(c)=0$ we have1800\begin{align*}1801q(c)1802&=1803\sum_{k=0}^{m-1} c_k c_{m+k}1804=18050.1806\end{align*}1807Let $K := \{ k \mid c_k c_{m+k} = 1 \}$.1808Then we must have $\abs{K} = 2 r$ for some integer $r \geqslant 0$, i.e. $\abs{K}$ is even.1809We therefore arbitrarily group the elements of $K$ into pairs $(i_p, j_p)$ for $p=0,\ldots,r-1$,1810and define the matrix $T \in \F_2^{m \times m}$ by1811\begin{align*}1812T_{i,j}1813&:=1814\sum_{p=0}^{r-1} (\delta_{i,i_p} \delta_{j,j_p} + \delta_{i,j_p} \delta_{j,i_p}),1815\end{align*}1816so that1817\begin{align*}1818\begin{cases}1819T_{i_p,j_p}1820=1821T_{j_p,i_p}1822=182311824&\text{for~} p \in \{0,\ldots,r-1\},1825\\1826T_{i,j} = 01827&\text{otherwise.}1828\end{cases}1829\end{align*}1830Since the $r$ pairs $(i_p, j_p)$ partition the set $K$,1831the matrix $T$ has at most one non-zero in each row and column.18321833Recalling that1834\begin{align*}1835(T^2)_{i,j}1836&=1837\sum_{k=0}^{m-1} T_{i,k} T_{k,j},1838\end{align*}1839we see that the general term $T_{i,k} T_{k,j}$ of this sum is non-zero only if either1840\begin{align*}1841\begin{cases}1842i = j = i_p,&\text{and}\ k=j_p,\ \text{or}1843\\1844i = j = j_p,&\text{and}\ k=i_p,1845\end{cases}1846\end{align*}1847for some $p \in \{0,\ldots,r-1\}$, with all $2r$ of these cases being mutually exclusive.1848So $T^2$ is diagonal with $2r$ non-zeros at the elements of $K$.18491850But $C_{1,1} C_{0,0}$ is diagonal, and $(C_{1,1} C_{0,0})_{i,i} = c_{m+i} c_i$.1851Therefore1852\begin{align}1853T^2 &= C_{1,1} C_{0,0}.1854\label{eq-t-2}1855\end{align}18561857Now, let $Z_{0,0}=Z_{1,1}=T$. Then $S_{0,0} = T + C_{0,0}$, $S_{1,1} = T + C_{1,1}$, and1858\begin{align*}1859S_{1,1} S_{0,0}1860&=1861(T + C_{1,1})(T + C_{0,0})1862=1863T^2 + T C_{0,0} + C_{1,1} T + C_{1,1} C_{0,0}1864\\1865&=1866T C_{0,0} + C_{1,1} T,1867\end{align*}1868where in the last step, we have used \eqref{eq-t-2}.18691870Now,1871\begin{align*}1872(T C_{0,0} + C_{1,1} T)_{i,j}1873&=1874\sum_{k=0}^{m-1} T_{i,k} (C_{0,0})_{k,j} + (C_{1,1})_{i,k} T_{k,j}1875\\1876&=1877T_{i,j} (C_{0,0})_{j,j} + (C_{1,1})_{i,i} T_{i,j}1878\\1879&=1880T_{i,j} \left( c_j + c_{m+i} \right).1881\end{align*}1882As above, $T_{i,j}$ is non-zero only when $(i,j)=(i_p,j_p)$ or $(i,j)=(j_p,i_p)$1883for some $p \in \{0,\ldots,r-1\}$, but in all those cases $c_j=c_{m+j}=1$.18841885Therefore1886\begin{align*}1887S_{1,1} S_{0,0} &= T C_{0,0} + C_{1,1} T = 0.1888\end{align*}1889Similarly, $S_{0,0} S_{1,1} = 0$, and therefore1890\begin{align*}1891A^21892&=1893\left[1894\begin{array}{cc}1895I & S_{1,1}1896\\1897S_{0,0} & I1898\end{array}1899\right]1900\left[1901\begin{array}{cc}1902I & S_{1,1}1903\\1904S_{0,0} & I1905\end{array}1906\right]1907\\1908&=1909\left[1910\begin{array}{cc}1911I + S_{1,1} S_{0,0} & S_{1,1} + S_{1,1}1912\\1913S_{0,0} + S_{0,0} & I + S_{0,0} S_{1,1}1914\end{array}1915\right]1916=1917\left[1918\begin{array}{cc}1919I & 01920\\19210 & I1922\end{array}1923\right].1924\end{align*}19251926We have therefore shown that1927\begin{align}1928A1929&:=1930\left[1931\begin{array}{cc}1932I & T + C_{1,1}1933\\1934T + C_{0,0} & I1935\end{array}1936\right],1937\quad1938S1939:=1940\left[1941\begin{array}{cc}1942T + C_{0,0} & 01943\\19440 & T + C_{1,1}1945\end{array}1946\right]1947\label{eq-a-s-def}1948\end{align}1949is a solution to $A^T L A = L + S$ with $A \in GL(2 m, 2)$.19501951Finally, given $c$ with $q(c)=0$, the matrix $A$ as defined by \eqref{eq-a-s-def} is such that1952$q(A x) = q(x) + \langle c, x \rangle$.1953\end{proof}1954% \end{frame}19551956\begin{Lemma}1957\label{lm-notes-6}1958For $k \in \{0,\ldots,m-1\}$ define $e^{(k)}$ by1959\begin{align}1960e_i^{(k)} &:= \delta_{i,k} + \delta_{i,m+k}1961\label{eq-e-def}1962\end{align}1963for $i \in \{0,\ldots,2 m - 1\}$.19641965Let $h(x) := q(x) + \langle e^{(0)}, x \rangle$, where $q$ is defined as per Lemma \ref{lm-notes-5}.1966Then for any $c'$ such that $q(c')=1$, there exists $B \in GL(2 m, 2)$ such that1967\begin{align}1968h(B x) &= q(x) + \langle c',x \rangle.1969\label{eq-h-B-x}1970\end{align}1971\end{Lemma}19721973\begin{proof}1974Let $K'=\{k \mid c'_k c'_{m+k} = 1\}$. Since $q(c')=1$, $\abs{K'}$ is odd.1975Choose any $\ell \in K'$, and let $c := c' + e^{(\ell)}$.1976Then $c_{\ell} = c_{m+\ell} = 0$ and $q(c)=0$.19771978Now let $h^{(\ell)}(x) := q(x) + \langle e^{(\ell)}, x \rangle$.1979We calculate1980\begin{align*}1981h^{(\ell)}(A x)1982&=1983q(A x) + \langle e^{(\ell)}, A x \rangle1984=1985q(x) + \langle c, x \rangle + \langle A^T e^{(\ell)}, x \rangle1986\\1987&=1988q(x) + \langle c + A^T e^{(\ell)}, x \rangle1989\end{align*}1990for $A$ given by the proof of Lemma \ref{lm-notes-4}.19911992If we let $K := \{ k \mid c_k c_{m+k} = 1 \}$, we see that $K = K' \setminus \{\ell\}$.1993Applying the other definitions and techniques used in the proof of Lemma \ref{lm-notes-4},1994we see that since $c_{\ell} = c_{m+\ell} = 0$ and $K$ does not contain $\ell$,1995column $\ell$ of each of $S_{0,0} := T + C_{0,0}$1996and $S_{1,1} := T + C_{1,1}$ is $0$, and therefore columns $\ell$ and $m + \ell$ of1997\begin{align*}1998A^T + I1999&:=2000\left[2001\begin{array}{cc}2002I & T + C_{0,0}2003\\2004T + C_{1,1} & I2005\end{array}2006\right]2007\end{align*}2008are both $0$.2009Therefore $A^T e^{(\ell)} = e^{(\ell)}$, and therefore2010\begin{align*}2011h^{(\ell)}(A x)2012&=2013q(x) + \langle c', x \rangle.2014\end{align*}20152016\end{proof}20172018\begin{Lemma}2019\label{lm-notes-6-b}2020For distinct $k,\ell \in \{0,\ldots,m-1\}$ let $e^{(k)}, e^{(\ell)}$ be defined as per Lemma2021\ref{lm-notes-6}.2022Let $h(x) := q(x) + \langle e^{(k)}, x \rangle$, where $q$ is defined as per Lemma \ref{lm-notes-5}.2023Then there exists $A \in GL(2 m, 2)$ such that2024\begin{align}2025h(A x)2026&=2027q(x) + \langle e^{(\ell)},x \rangle.2028\end{align}2029\end{Lemma}20302031\begin{proof}2032The matrix $A$ is the permutation matrix for the the permutation $(k\ \ell)(m+k\ m+\ell)$ (defined2033using cycle notation.)2034\end{proof}20352036\begin{Lemma}2037\label{lm-notes-7}2038Let $q$ be defined as per Lemma \ref{lm-notes-5}.2039Then for all $c, c' \in Z_2^{2 m}$ with $q(c)=q(c')=1$, there exists $A \in GL(2 m, 2)$ such that2040if $h(x) := q(x) + \langle c, x \rangle$, then2041\begin{align*}2042h(A x) &= q(x) + \langle c', x \rangle.2043\end{align*}2044\end{Lemma}20452046\begin{proof}2047This is a consequence of Lemmas \ref{lm-notes-6} and \ref{lm-notes-6-b}.2048\end{proof}20492050\begin{proofof}{Theorem \ref{th-Quadratic-Classes}}2051It is well known that all quadratic bent functions are contained in one Extended Affine equivalence2052class.2053As a consequence of Corollary \ref{corr-Affine-Translate-Cayley}, without loss of generality, we need2054only examine2055the Extended Translation equivalence class of the quadratic function $q$ as defined in Lemma2056\ref{lm-notes-5}.20572058As a result of Lemma \ref{lm-notes-5}, we actually need only examine functions of the form2059$f(x) = q(x) + \langle c,x \rangle$2060for some $c \in \F_2^{2m}$.2061Lemma \ref{lm-notes-4} implies that all such functions for which $q(c)=0$ are Cayley equivalent to2062$q$.2063Lemma \ref{lm-notes-7} implies that any two such functions $q(x) + \langle c, x \rangle$ and $q(x)\, +2064\langle c', x \rangle$2065with $q(c)=q(c')=1$ are Cayley equivalent to each other.20662067The functions where $q(c)=0$ are not Cayley equivalent to the functions where $q(c)=1$ because2068Lemma \ref{lm-notes-9b} implies that2069\begin{align*}2070\weightclass{x \mapsto q(x) + \langle c,x \rangle}2071&=2072\dual{q}(c) = q(c),2073\end{align*}2074since $q$ is self-dual.2075\end{proofof}20762077%\newpage20782079\section{Computational results for low dimensions}2080\label{sec-Empirical}2081This section lists some properties of bent functions and their extended affine (EA) classes, extended translation (ET) classes, and extended Cayley classes2082that have been computed for 2, 4, 6 and 8 dimensions.2083The computations were made using Sage \cite{SageMath7517} and CoCalc \cite{CoCalc}.2084Larger scale computations, involving millions of ET classes,2085were conducted on the Raijin supercomputer of the National Computational Infrastructure.2086Sage and Python code for these computations are available on GitHub \cite{Leo16GitHub} and CoCalc \cite{Leo17CoCalc}.2087Some CoCalc worksheets also illustrate these and related computations \cite{Leo17CoCalc}.2088The Sage and Python code is briefly described in Section \ref{sec-Code}.20892090%\newpage2091In the tables below, each bent function is defined by its algebraic normal form, and each Cayley class is described by2092its number within the ET class of the bent function (from 0, in the order in which Sage identified non-isomorphic graphs),2093followed by three properties of the Cayley graph: its parameters as a strongly regular graph,2094the 2-rank of its adjacency matrix \cite{Brov92}, and its clique polynomial \cite{HoeL94}.2095The 2-rank is included for comparison with Tonchev's tables of 2-weight codes in 6 dimensions \cite{Ton96uniformly,Ton07codes}.2096The clique polynomial is included for interest's sake, and to illustrate the variety of strongly regular graphs2097that exist with the same parameters, even for low dimensions.20982099The plots below are produced by the function \texttt{sage.}\texttt{plot.}\texttt{matrix\_plot},2100with \texttt{gist\_stern} as the \texttt{colormap}.2101Thus the smallest number is coloured black and the largest number is coloured white.21022103The plotted matrices all contain non-negative integers.2104The weight class matrices are defined by Definition~\ref{def-weight-class-matrix}, and are $\{0,1\}$ matrices,2105so their matrix plots are therefore black and white, with black representing 0 and white representing 1.2106The other matrices record the number of the Cayley class within the2107ET class, starting from 0, as per corresponding table of extended Cayley classes.21082109Some highlights of the computational results include:2110\begin{enumerate}2111\item Verification of Theorem \ref{th-Quadratic-Classes}: the quadratic bent functions have two Cayley classes2112corresponding to the two weight classes.2113\item In 6 dimensions, identification of the Cayley classes corresponding to2114Tonchev's tables of 2-weight codes \cite{Ton96uniformly,Ton07codes}.2115%See also the paper by Jungnickel and Tonchev \cite{JunT1992}.2116\item In 6 and 8 dimensions, extended Cayley equivalence between a quadratic bent function2117and a bent function of degree 3.2118In each case, the isomorphism between Cayley graphs is not a linear function on $\F_2^{2m}$.2119\item In 8 dimensions, the result that two of Braeken's extended affine classes of bent functions of degree2120at most 3 \cite{Bra06thesis, Tok15bent} are actually the same class.2121Thus the list only contains 9 distinct classes and not 10.2122\item In 8 dimensions, 8 of the 256 bent functions used for the S-boxes of the CAST-128 cypher \cite{RFC2144,Ada97}2123are exceptional in the sense that,2124for each of these 8 bent functions, the $65\,536$ bent functions in the extended translation class,2125and their $65\,536$ duals do not yield $131\,072$ distinct Cayley graphs.2126In contrast, for the remaining 248 bent functions, these $131\,072$ Cayley graphs are all non-isomorphic.2127\end{enumerate}21282129% \begin{frame}2130\subsection{Bent functions in 2 dimensions}2131%2132The bent functions on $\F_2^2$ consist of one EA class, containing the ET class: $[f_{2,1}]$2133where $f_{2,1}(x) := x_0 x_1$ is self dual.2134The ET class contains two extended Cayley classes as per Table~\ref{tab-c2_1_EC_classes}.2135Note that the Cayley graph for class 1 is $K_4$, which is not considered to be strongly regular, by convention.2136\begin{table}[!bhpt] % [H]2137\small{2138\begin{align*}2139\def\arraystretch{1.2}2140\begin{array}{|cccl|}2141\hline2142\text{Class} &2143\text{Parameters} &2144\text{2-rank} &2145\text{Clique polynomial}2146\\2147\hline21480 &2149(4, 1, 0, 0) &21504 &2151\begin{array}{l}21522t^{2} + 4t + 12153\end{array}2154\\21551 &2156K_4 &21574 &2158\begin{array}{l}2159t^{4} + 4t^{3} + 6t^{2} + 4t + 12160\end{array}2161\\2162\hline2163\end{array}2164\end{align*}2165}2166\caption{$[f_{2,1}]$ extended Cayley classes.}2167\label{tab-c2_1_EC_classes}2168\end{table}21692170\begin{figure}[!ht]2171\centering2172\begin{minipage}{.48\textwidth}2173\centering2174\includegraphics[width=.9\linewidth]{c2_1_weight_class_matrix.png}2175\captionof{figure}{$[f_{2,1}]$: weight classes. ~~~~\\~~~~}2176\label{fig:c2_1_weight_class_matrix}2177\end{minipage}%2178~~~~2179\begin{minipage}{.48\textwidth}2180\centering2181\includegraphics[width=.9\linewidth]{c2_1_bent_cayley_graph_index_matrix.png}2182\captionof{figure}{$[f_{2,1}]$: extended Cayley classes.}2183\label{fig:c2_1_bent_cayley_graph_index_matrix}2184\end{minipage}2185\end{figure}2186As expected from Theorem~\ref{th-Quadratic-Classes},2187the two extended Cayley classes correspond to the two weight classes,2188as shown in Figures~\ref{fig:c2_1_weight_class_matrix} and~\ref{fig:c2_1_bent_cayley_graph_index_matrix}.21892190% \end{frame}2191% \begin{frame}2192\subsection{Bent functions in 4 dimensions}2193%2194The bent functions on $\F_2^4$ consist of one EA class, containing the ET class $[f_{4,1}]$ where2195$f_{4,1}(x) := x_0 x_1 + x_2 x_3$ is self dual.2196The ET class contains two extended Cayley classes as per Table~\ref{tab-c4_1_EC_classes}.2197\begin{table}[!bhpt] % [H]2198\small{2199\begin{align*}2200\def\arraystretch{1.2}2201\begin{array}{|cccl|}2202\hline2203\text{Class} &2204\text{Parameters} &2205\text{2-rank} &2206\text{Clique polynomial}2207\\2208\hline22090 &2210(16, 6, 2, 2) &22116 &2212\begin{array}{l}22138t^{4} + 32t^{3} + 48t^{2} + 16t + 12214\end{array}2215\\22161 &2217(16, 10, 6, 6) &22186 &2219\begin{array}{l}222016t^{5} + 120t^{4} + 160t^{3}2221\,+2222\\222380t^{2} + 16t + 12224\end{array}2225\\2226\hline2227\end{array}2228\end{align*}2229}2230\caption{$[f_{4,1}]$ extended Cayley classes.}2231\label{tab-c4_1_EC_classes}2232\end{table}22332234% The Cayley graphs for classes 1 and 2 are isomorphic to those those obtained from the2235% projective two-weight codes listed in Table~\ref{tab-c4_1_codes}2236%2237% \begin{table}[!bhpt] % [H]2238% \begin{align*}2239% \begin{array}{|ccc|}2240% \hline2241% \text{Class} &2242% \text{Parameters} & \text{Generator matrix}2243% \\2244% \hline2245% 1 &2246% [6, 4, 2] &2247% \left[2248% \begin{array}{cccccc}2249% 0 & 0 & 1 & 1 & 1 & 12250% \\2251% 1 & 0 & 0 & 1 & 1 & 12252% \\2253% 1 & 1 & 1 & 1 & 0 & 02254% \\2255% 0 & 1 & 1 & 1 & 1 & 02256% \end{array}2257% \right]2258% \\2259% 2 &2260% [5, 4, 2] &2261% \left[2262% \begin{array}{ccccc}2263% 1 & 1 & 0 & 0 & 02264% \\2265% 0 & 1 & 1 & 0 & 02266% \\2267% 0 & 0 & 0 & 1 & 12268% \\2269% 1 & 0 & 0 & 0 & 12270% \end{array}2271% \right]2272% \\2273% \hline2274% \end{array}2275% \end{align*}2276% \caption{$f_{4,1}$ Two-weight projective codes}2277% \label{tab-c4_1_codes}2278% \end{table}22792280\begin{figure}[!bhpt] % [H]2281\centering2282\begin{minipage}{.48\textwidth}2283\centering2284\includegraphics[width=.9\linewidth]{c4_1_weight_class_matrix.png}2285\captionof{figure}{$[f_{4,1}]$: weight classes. ~~~~\\~~~~}2286\label{fig:c4_1_weight_class_matrix}2287\end{minipage}%2288~~~~2289\begin{minipage}{.48\textwidth}2290\centering2291\includegraphics[width=.9\linewidth]{c4_1_bent_cayley_graph_index_matrix.png}2292\captionof{figure}{$[f_{4,1}]$: extended Cayley classes.}2293\label{fig:c4_1_bent_cayley_graph_index_matrix}2294\end{minipage}2295\end{figure}2296The two extended Cayley classes correspond to the two weight classes,2297as shown in Figures~\ref{fig:c4_1_weight_class_matrix} and~\ref{fig:c4_1_bent_cayley_graph_index_matrix}.22982299% \end{frame}2300\newpage2301% \begin{frame}2302\subsection{Bent functions in 6 dimensions}2303\paragraph*{Extended affine classes.}2304%2305The bent functions on $\F_2^6$ consist of four2306EA classes, containing the ET classes as listed in Table~\ref{tab-c6_ET_classes}2307\cite[p. 303]{Rot76} \cite[Section 7.2]{Tok15bent}.2308\begin{table}[!bhpt] % [H]2309\small{2310\begin{align*}2311\def\arraystretch{1.2}2312\begin{array}{|cl|}2313\hline2314\text{Class} &2315\text{Representative}2316\\2317\hline2318\,[f_{6,1}] & f_{6,1} :=2319\begin{array}{l}2320x_{0} x_{1} + x_{2} x_{3} + x_{4} x_{5}2321\end{array}2322\\2323\,[f_{6,2}] & f_{6,2} :=2324\begin{array}{l}2325x_{0} x_{1} x_{2} + x_{0} x_{3} + x_{1} x_{4} + x_{2} x_{5}2326\end{array}2327\\2328\,[f_{6,3}] & f_{6,3} :=2329\begin{array}{l}2330x_{0} x_{1} x_{2} + x_{0} x_{1} + x_{0} x_{3} + x_{1} x_{3} x_{4} + x_{1} x_{5}\, +2331\\2332x_{2} x_{4} + x_{3} x_{4}2333\end{array}2334\\2335\,[f_{6,4}] & f_{6,4} :=2336\begin{array}{l}2337x_{0} x_{1} x_{2} + x_{0} x_{3} + x_{1} x_{3} x_{4} + x_{1} x_{5} + x_{2} x_{3} x_{5}\, +2338\\2339x_{2} x_{3} + x_{2} x_{4} + x_{2} x_{5} + x_{3} x_{4} + x_{3} x_{5}2340\end{array}2341\\2342\hline2343\end{array}2344\end{align*}2345}2346\caption{6 dimensions: ET classes.}2347\label{tab-c6_ET_classes}2348\end{table}23492350%\newpage2351In 1996, Tonchev classified the binary projective two-weight $[27,21,3]$ and $[35,6,16]$ codes2352listing them in Tables 1 and 2, respectively, of his paper \cite{Ton96uniformly}.2353These tables are repeated as Tables 1.155 and 1.156 in Chapter VII.1 of the Handbook of2354Combinatorial Designs, Second Edition \cite{Ton07codes},2355with a different numbering.2356For each of the codes listed in these two tables, the characteristics of the corresponding2357strongly regular graph is also listed.23582359In the classification given below, the Cayley graph of each Cayley class is matched by isomorphism2360with a strongly regular graph corresponding to one2361or more of Tonchev's projective two-weight codes, or the complement of such a graph.2362Tonchev's strongly regular graphs were checked using the function2363\verb!strongly_regular_from_two_weight_code!, which uses the smaller of the two weights to create the graph2364\cite{SageMath7517}.2365% \end{frame}2366% \begin{frame}2367%2368\paragraph*{ET class $[f_{6,1}]$.}2369%2370This is the ET class of the bent function2371$f_{6,1}(x) := x_0 x_1 + x_2 x_3 + x_4 x_5.$2372This function is quadratic and self-dual.23732374The ET class contains two extended Cayley classes as per Table~\ref{tab-c6_1_EC_classes}.23752376\begin{table}[!bhpt] % [H]2377%2378\small{}2379\begin{align*}2380\def\arraystretch{1.2}2381\begin{array}{|cccl|}2382\hline2383\text{Class} &2384\text{Parameters} &2385\text{2-rank} &2386\text{Clique polynomial}2387\\2388\hline23890 &2390(64, 28, 12, 12) &23918 &2392\begin{array}{l}239364t^{8} + 512t^{7} + 1792t^{6} + 3584t^{5}2394\,+2395\\23965376t^{4} + 3584t^{3} + 896t^{2} + 64t + 12397\end{array}2398\\23991 &2400(64, 36, 20, 20) &24018 &2402\begin{array}{l}24032304t^{6} + 13824t^{5} + 19200t^{4} + 7680t^{3}2404\,+2405\\24061152t^{2} + 64t + 12407\end{array}2408\\2409\hline2410\end{array}2411\end{align*}2412%2413\caption{$[f_{6,1}]$ extended Cayley classes.}2414\label{tab-c6_1_EC_classes}2415\end{table}24162417The Cayley graphs for classes 0 and 1 are isomorphic to those those obtained from2418Tonchev's projective two-weight codes \cite{Ton07codes} as per Table~\ref{tab-c6_1_codes}.24192420\begin{table}[!bhpt] % [H]2421\small{2422\begin{align*}2423\def\arraystretch{1.2}2424\begin{array}{|ccl|}2425\hline2426\text{Class} &2427\text{Parameters} & \text{Reference}2428\\2429\hline24300 & [35,6,16] & \text{Table 1.156 1, 2 (complement)}2431\\24321 & [27,6,12] & \text{Table 1.155 1 }2433\\2434\hline2435\end{array}2436\end{align*}2437}2438\caption{$[f_{6,1}]$ Two-weight projective codes.}2439\label{tab-c6_1_codes}2440\end{table}24412442%\slidecite{Tonchev 1996, 2006}2443% \end{frame}2444% \begin{frame}24452446The two extended Cayley classes correspond to the two weight classes,2447as shown in Figures~\ref{fig:c6_1_weight_class_matrix} and~\ref{fig:c6_1_bent_cayley_graph_index_matrix}.24482449\begin{figure}[!bhpt] % [H]2450\centering2451\begin{minipage}{.48\textwidth}2452\centering2453\includegraphics[width=.9\linewidth]{c6_1_weight_class_matrix.png}2454\captionof{figure}{$[f_{6,1}]$: weight classes. ~~~~\\~~~~}2455\label{fig:c6_1_weight_class_matrix}2456\end{minipage}%2457~~~~2458\begin{minipage}{.48\textwidth}2459\centering2460\includegraphics[width=.9\linewidth]{c6_1_bent_cayley_graph_index_matrix.png}2461\captionof{figure}{$[f_{6,1}]$: extended Cayley classes.}2462\label{fig:c6_1_bent_cayley_graph_index_matrix}2463\end{minipage}2464\end{figure}2465%\newpage2466Remark: The sequence of Figures~\ref{fig:c2_1_weight_class_matrix}, \ref{fig:c4_1_weight_class_matrix},2467and \ref{fig:c6_1_weight_class_matrix} displays a fractal-like self-similar quality.24682469\paragraph*{ET class $[f_{6,2}]$.}2470%2471This is the ET class of the bent function2472$f_{6,2}(x) := x_{0} x_{1} x_{2} + x_{0} x_{3} + x_{1} x_{4} + x_{2} x_{5}$.24732474The ET class contains three extended Cayley classes as per Table~\ref{tab-c6_2_EC_classes}.24752476\begin{table}[!bhpt] % [H]2477%2478\small{}2479\begin{align*}2480\def\arraystretch{1.2}2481\begin{array}{|cccl|}2482\hline2483\text{Class} &2484\text{Parameters} &2485\text{2-rank} &2486\text{Clique polynomial}2487\\2488\hline24890 &2490(64, 28, 12, 12) &24918 &2492\begin{array}{l}249364t^{8} + 512t^{7} + 1792t^{6} + 3584t^{5}2494\,+2495\\24965376t^{4} + 3584t^{3} + 896t^{2} + 64t + 12497\end{array}2498\\24991 &2500(64, 28, 12, 12) &25018 &2502\begin{array}{l}2503256t^{6} + 1536t^{5} + 4352t^{4} + 3584t^{3}2504\,+2505\\2506896t^{2} + 64t + 12507\end{array}2508\\25092 &2510(64, 36, 20, 20) &25118 &2512\begin{array}{l}2513192t^{8} + 1536t^{7} + 8960t^{6} + 19968t^{5}2514\,+2515\\251620224t^{4} + 7680t^{3} + 1152t^{2} + 64t + 12517\end{array}2518\\2519\hline2520\end{array}2521\end{align*}2522%2523\caption{$[f_{6,2}]$ extended Cayley classes.}2524\label{tab-c6_2_EC_classes}2525\end{table}25262527The Cayley graph for class 0 is isomorphic to graph 0 of ET class $[f_{6,1}]$,2528This reflects the fact that $f_{6,1} \equiv f_{6,2}$, even though these two functions are not2529EA equivalent.2530This is therefore an example of an isomorphism between Cayley graphs of bent functions on2531$\F_2^6$ that is not a linear function.25322533The Cayley graph for class 0 is also isomorphic to the complement of Royle's $(64,35,18,20)$ strongly regular graph $X$2534\cite{Roy08normal}.25352536%2537%~2538%2539%\slidecite{Royle 2008}2540% \end{frame}2541% \begin{frame}2542%\paragraph*{ET class $[f_{6,2}]$: two-weight codes.}25432544The Cayley graphs for classes 0 to 2 are isomorphic to those those obtained from2545Tonchev's projective two-weight codes \cite{Ton07codes} as per Table~\ref{tab-c6_2_codes}.25462547\begin{table}[!bhpt] % [H]2548\small{2549\begin{align*}2550\def\arraystretch{1.2}2551\begin{array}{|ccl|}2552\hline2553\text{Class} &2554\text{Parameters} & \text{Reference}2555\\2556\hline25570 & [35,6,16] & \text{Table 1.156 1, 2 (complement)}2558\\25591 & [35,6,16] & \text{Table 1.156 3 (complement)}2560\\25612 & [27,6,12] & \text{Table 1.155 2 }2562\\2563\hline2564\end{array}2565\end{align*}2566}2567\caption{$[f_{6,2}]$ Two-weight projective codes.}2568\label{tab-c6_2_codes}2569\end{table}2570%\newpage25712572The three extended Cayley classes are distributed between the two weight classes,2573as shown in Figures~\ref{fig:c6_2_weight_class_matrix} and~\ref{fig:c6_2_bent_cayley_graph_index_matrix}.25742575\begin{figure}[!bhpt] % [H]2576\centering2577\begin{minipage}{.48\textwidth}2578\centering2579\includegraphics[width=.9\linewidth]{c6_2_weight_class_matrix.png}2580\captionof{figure}{$[f_{6,2}]$: weight classes. ~~~~\\~~~~}2581\label{fig:c6_2_weight_class_matrix}2582\end{minipage}%2583~~~~2584\begin{minipage}{.48\textwidth}2585\centering2586\includegraphics[width=.9\linewidth]{c6_2_bent_cayley_graph_index_matrix.png}2587\captionof{figure}{$[f_{6,2}]$: extended Cayley classes.}2588\label{fig:c6_2_bent_cayley_graph_index_matrix}2589\end{minipage}2590\end{figure}25912592%\slidecite{Tonchev 1996, 2006}2593% \end{frame}2594%%\newpage2595% \begin{frame}2596%2597\paragraph*{ET class $[f_{6,3}]$.}2598%2599This is the ET class of the bent function2600\begin{align*}2601f_{6,3}(x) &= x_{0} x_{1} x_{2} + x_{0} x_{1} + x_{0} x_{3} + x_{1} x_{3} x_{4}2602\\2603&+ x_{1} x_{5} + x_{2} x_{4} + x_{3} x_{4}.2604\end{align*}26052606The ET class contains four extended Cayley classes as per Table~\ref{tab-c6_3_EC_classes}.26072608\begin{table}[!bhpt] % [H]2609%2610\small{}2611\begin{align*}2612\def\arraystretch{1.2}2613\begin{array}{|cccl|}2614\hline2615\text{Class} &2616\text{Parameters} &2617\text{2-rank} &2618\text{Clique polynomial}2619\\2620\hline26210 &2622(64, 28, 12, 12) &262312 &2624\begin{array}{l}262532t^{8} + 256t^{7} + 896t^{6} + 2048t^{5} + 4608t^{4}2626\,+2627\\26283584t^{3} + 896t^{2} + 64t + 12629\end{array}2630\\26311 &2632(64, 36, 20, 20) &263312 &2634\begin{array}{l}2635160t^{8} + 1280t^{7} + 9344t^{6} + 21504t^{5}2636\,+2637\\263820480t^{4} + 7680t^{3} + 1152t^{2} + 64t + 12639\end{array}2640\\26412 &2642(64, 28, 12, 12) &264312 &2644\begin{array}{l}264564t^{6} + 1024t^{5} + 4096t^{4} + 3584t^{3}2646\,+2647\\2648896t^{2} + 64t + 12649\end{array}2650\\26513 &2652(64, 36, 20, 20) &265312 &2654\begin{array}{l}2655160t^{8} + 1664t^{7} + 9792t^{6} + 21504t^{5}2656\,+2657\\265820480t^{4} + 7680t^{3} + 1152t^{2} + 64t + 12659\end{array}2660\\2661\hline2662\end{array}2663\end{align*}2664%2665\caption{$[f_{6,3}]$ extended Cayley classes.}2666\label{tab-c6_3_EC_classes}2667\end{table}26682669The Cayley graphs for classes 0 to 3 are isomorphic to those those obtained from2670Tonchev's projective two-weight codes \cite{Ton07codes} as per Table~\ref{tab-c6_3_codes}.26712672\begin{table}[!bhpt] % [H]2673\small{2674\begin{align*}2675\def\arraystretch{1.2}2676\begin{array}{|ccl|}2677\hline2678\text{Class} &2679\text{Parameters} & \text{Reference}2680\\2681\hline26820 & [35,6,16] & \text{Table 1.156 4 (complement)}2683\\26841 & [27,6,12] & \text{Table 1.155 3 }2685\\26862 & [35,6,16] & \text{Table 1.156 5 (complement)}2687\\26883 & [27,6,12] & \text{Table 1.155 4 }2689\\2690\hline2691\end{array}2692\end{align*}2693}2694\caption{$[f_{6,3}]$ Two-weight projective codes.}2695\label{tab-c6_3_codes}2696\end{table}26972698The four extended Cayley classes are distributed between the two weight classes,2699as shown in Figures~\ref{fig:c6_3_weight_class_matrix} and~\ref{fig:c6_3_bent_cayley_graph_index_matrix}.27002701\begin{figure}[!ht] % [H]2702\centering2703\begin{minipage}{.48\textwidth}2704\centering2705\includegraphics[width=.9\linewidth]{c6_3_weight_class_matrix.png}2706\captionof{figure}{$[f_{6,3}]$: weight classes. ~~~~\\~~~~}2707\label{fig:c6_3_weight_class_matrix}2708\end{minipage}%2709~~~~2710\begin{minipage}{.48\textwidth}2711\centering2712\includegraphics[width=.9\linewidth]{c6_3_bent_cayley_graph_index_matrix.png}2713\captionof{figure}{$[f_{6,3}]$: extended Cayley classes.}2714\label{fig:c6_3_bent_cayley_graph_index_matrix}2715\end{minipage}2716\end{figure}27172718%\slidecite{Tonchev 1996, 2006}2719% \end{frame}2720% \begin{frame}2721%\newpage2722\paragraph*{ET class $[f_{6,4}]$.}2723%2724This is the ET class of the bent function2725\begin{align*}2726f_{6,4}(x) &= x_{0} x_{1} x_{2} + x_{0} x_{3} + x_{1} x_{3} x_{4} + x_{1} x_{5} + x_{2} x_{3} x_{5}2727\\2728&+ x_{2} x_{3} + x_{2} x_{4} + x_{2} x_{5} + x_{3} x_{4} + x_{3} x_{5}.2729\end{align*}27302731The ET class contains three extended Cayley classes as per Table~\ref{tab-c6_4_EC_classes}.27322733\begin{table}[!bhpt] % [H]2734%2735\small{}2736\begin{align*}2737\def\arraystretch{1.2}2738\begin{array}{|cccl|}2739\hline2740\text{Class} &2741\text{Parameters} &2742\text{2-rank} &2743\text{Clique polynomial}2744\\2745\hline27460 &2747(64, 28, 12, 12) &274814 &2749\begin{array}{l}275032t^{8} + 256t^{7} + 896t^{6} + 1792t^{5} + 4480t^{4}2751\,+2752\\27533584t^{3} + 896t^{2} + 64t + 12754\end{array}2755\\27561 &2757(64, 28, 12, 12) &275814 &2759\begin{array}{l}276016t^{8} + 128t^{7} + 448t^{6} + 1280t^{5} + 4224t^{4}2761\,+2762\\27633584t^{3} + 896t^{2} + 64t + 12764\end{array}2765\\27662 &2767(64, 36, 20, 20) &276814 &2769\begin{array}{l}2770176t^{8} + 1408t^{7} + 9664t^{6} + 22272t^{5}2771\,+2772\\277320608t^{4} + 7680t^{3} + 1152t^{2} + 64t + 12774\end{array}2775\\2776\hline2777\end{array}2778\end{align*}2779\caption{$[f_{6,4}]$ extended Cayley classes.}2780\label{tab-c6_4_EC_classes}2781\end{table}27822783The Cayley graphs for classes 0 to 2 are isomorphic to those those obtained from2784Tonchev's projective two-weight codes \cite{Ton07codes} as per Table~\ref{tab-c6_4_codes}.27852786\begin{table}[!bhpt] % [H]2787\small{2788\begin{align*}2789\def\arraystretch{1.2}2790\begin{array}{|ccl|}2791\hline2792\text{Class} &2793\text{Parameters} & \text{Reference}2794\\2795\hline27960 & [35,6,16] & \text{Table 1.156 7 (complement)}2797\\27981 & [35,6,16] & \text{Table 1.156 6 (complement)}2799\\28002 & [27,6,12] & \text{Table 1.155 5 }2801\\2802\hline2803\end{array}2804\end{align*}2805}2806\caption{$[f_{6,4}]$ Two-weight projective codes.}2807\label{tab-c6_4_codes}2808\end{table}28092810%\slidecite{Tonchev 1996, 2006}2811% \end{frame}2812% \begin{frame}28132814The three extended Cayley classes are distributed between the two weight classes,2815as shown in Figures~\ref{fig:c6_4_weight_class_matrix} and~\ref{fig:c6_4_bent_cayley_graph_index_matrix}.28162817\begin{figure}[!hpt] % [H]2818\centering2819\begin{minipage}{.48\textwidth}2820\centering2821\includegraphics[width=.9\linewidth]{c6_4_weight_class_matrix.png}2822\captionof{figure}{$[f_{6,4}]$: weight classes. ~~~~\\~~~~}2823\label{fig:c6_4_weight_class_matrix}2824\end{minipage}%2825~~~~2826\begin{minipage}{.48\textwidth}2827\centering2828\includegraphics[width=.9\linewidth]{c6_4_bent_cayley_graph_index_matrix.png}2829\captionof{figure}{$[f_{6,4}]$: extended Cayley classes.}2830\label{fig:c6_4_bent_cayley_graph_index_matrix}2831\end{minipage}2832\end{figure}2833\newpage2834\subsection{Bent functions in 8 dimensions}28352836There are2837$99\,270\,589\,265\,934\,370\,305\,785\,861\,242\,880 \approx 2^{106}$ bent functions in 8 dimensions,2838according to Langevin and Leander \cite{LanL11counting}.2839%2840%~2841%2842The number of EA classes has not yet been published,2843let alone a list of representative bent functions.2844The lists of EA classes of bent functions that have so far been published include those2845for the bent functions of degree at most 3 \cite[Section 5.5.2]{Bra06thesis} \cite[Section 7.3]{Tok15bent},2846and the partial spread bent functions \cite{Lan10psf,LanH11counting}.2847The bent functions used in the S-boxes of the CAST-128 encryption algorithm \cite{Ada97,RFC2144}2848are also representatives of disjoint EA classes.2849%\newpage2850\paragraph*{Extended affine classes of degree at most 3.}2851%2852According to a list contained in Braeken's PhD thesis \cite[Section 5.5.2]{Bra06thesis},2853and repeated in Tokareva's table \cite[Section 7.3]{Tok15bent},2854the bent functions on $\F_2^8$, of degree at most 3, consist of 102855EA classes, whose representatives are listed in Table~\ref{tab-c8_ET_classes}.2856\begin{table}[!bhpt] % [H]2857\small{}2858\begin{align*}2859\def\arraystretch{1.2}2860\begin{array}{|cl|}2861\hline2862\text{Class} &2863\text{Representative}2864\\2865\hline2866\,[f_{ 8 , 1 }] & f_{ 8 , 1 } :=2867\begin{array}{l}2868x_{0} x_{1} + x_{2} x_{3} + x_{4} x_{5} + x_{6} x_{7}2869\end{array}2870\\2871\,[f_{ 8 , 2 }] & f_{ 8 , 2 } :=2872\begin{array}{l}2873x_{0} x_{1} x_{2} + x_{0} x_{3} + x_{1} x_{4} + x_{2} x_{5} + x_{6} x_{7}2874\end{array}2875\\2876\,[f_{ 8 , 3 }] & f_{ 8 , 3 } :=2877\begin{array}{l}2878x_{0} x_{1} x_{2} + x_{0} x_{6} + x_{1} x_{3} x_{4} + x_{1} x_{5} + x_{2} x_{3} + x_{4} x_{7}2879\end{array}2880\\2881\,[f_{ 8 , 4 }] & f_{ 8 , 4 } :=2882\begin{array}{l}2883x_{0} x_{1} x_{2} + x_{0} x_{2} + x_{0} x_{4} + x_{1} x_{3} x_{4} + x_{1} x_{5} + x_{2} x_{3}\, +2884x_{6} x_{7}2885\end{array}2886\\2887\,[f_{ 8 , 5 }] & f_{ 8 , 5 } :=2888\begin{array}{l}2889x_{0} x_{1} x_{2} + x_{0} x_{6} + x_{1} x_{3} x_{4} + x_{1} x_{4} + x_{1} x_{5} + x_{2} x_{3} x_{5}2890+ x_{2} x_{4} + x_{3} x_{7}2891\end{array}2892\\2893\,[f_{ 8 , 6 }] & f_{ 8 , 6 } :=2894\begin{array}{l}2895x_{0} x_{1} x_{2} + x_{0} x_{2} + x_{0} x_{3} + x_{1} x_{3} x_{4} + x_{1} x_{6} + x_{2} x_{3} x_{5}2896+ x_{2} x_{4} + x_{5} x_{7}2897\end{array}2898\\2899\,[f_{ 8 , 7 }] & f_{ 8 , 7 } :=2900\begin{array}{l}2901x_{0} x_{1} x_{2} + x_{0} x_{1} + x_{0} x_{2} + x_{0} x_{3} + x_{1} x_{3} x_{4} + x_{1} x_{4}\, +2902x_{1} x_{5}\, +2903\\2904x_{2} x_{3} x_{5} + x_{2} x_{4} + x_{6} x_{7}2905\end{array}2906\\2907\,[f_{ 8 , 8 }] & f_{ 8 , 8 } :=2908\begin{array}{l}2909x_{0} x_{1} x_{2} + x_{0} x_{5} + x_{1} x_{3} x_{4} + x_{1} x_{6} + x_{2} x_{3} x_{5} + x_{2} x_{4}2910+ x_{3} x_{7}2911\end{array}2912\\2913\,[f_{ 8 , 9 }] & f_{ 8 , 9 } :=2914\begin{array}{l}2915x_{0} x_{1} x_{6} + x_{0} x_{3} + x_{1} x_{4} + x_{2} x_{3} x_{6} + x_{2} x_{5} + x_{3} x_{4}\, +2916x_{4} x_{5} x_{6} + x_{6} x_{7}2917\end{array}2918\\2919\,[f_{ 8 , 10 }] & f_{ 8 , 10 } :=2920\begin{array}{l}2921x_{0} x_{1} x_{2} + x_{0} x_{3} x_{6} + x_{0} x_{4} + x_{0} x_{5} + x_{1} x_{3} x_{4} + x_{1} x_{6}2922+ x_{2} x_{3} x_{5}\, +2923\\2924x_{2} x_{4} + x_{3} x_{7}2925\end{array}2926\\2927\hline2928\end{array}2929\end{align*}2930\normalsize{}2931\caption{8 dimensions to degree 3: ET classes.}2932\label{tab-c8_ET_classes}2933\end{table}2934% \end{frame}2935We here examine the corresponding ET classes in detail.2936% \begin{frame}2937\newpage2938\paragraph*{ET class $[f_{8,1}]$.}2939%2940This is the ET class of the bent function2941\small{}2942\begin{align*}2943f_{ 8 , 1 } &=2944\begin{array}{l}2945x_{0} x_{1} + x_{2} x_{3} + x_{4} x_{5} + x_{6} x_{7}.2946\end{array}2947\end{align*}2948\normalsize{}2949This function is quadratic and self-dual.2950The ET class contains two extended Cayley classes as per Table~\ref{tab-c8_1_EC_classes}.29512952% f8_12953\begin{table}[!bhpt] % [H]2954%2955\small{}2956\begin{align*}2957\def\arraystretch{1.2}2958\begin{array}{|cccl|}2959\hline2960\text{Class} &2961\text{Parameters} &2962\text{2-rank} &2963\text{Clique polynomial}2964\\2965\hline29660 &2967(256, 120, 56, 56) &296810 &2969\begin{array}{l}2970245760t^{9} + 3317760t^{8} + 8847360t^{7}2971\,+2972\\297310321920t^{6} + 6193152t^{5} + 2007040t^{4}2974\,+2975\\2976286720t^{3} + 15360t^{2} + 256t + 12977\end{array}2978\\29791 &2980(256, 136, 72, 72) &298110 &2982\begin{array}{l}2983417792t^{8} + 3342336t^{7} + 11698176t^{6}2984\,+2985\\298611698176t^{5} + 3760128t^{4} + 417792t^{3}2987\,+2988\\298917408t^{2} + 256t + 12990\end{array}2991\\2992\hline2993\end{array}2994\end{align*}2995%2996\caption{$[f_{8,1}]$ extended Cayley classes.}2997\label{tab-c8_1_EC_classes}2998\end{table}29993000As expected from Theorem~\ref{th-Quadratic-Classes},3001the two extended Cayley classes correspond to the two weight classes,3002as shown in Figures~\ref{fig:c8_1_weight_class_matrix} and~\ref{fig:c8_1_bent_cayley_graph_index_matrix}.30033004Remark: The fractal-like self-similar quality of Figures~\ref{fig:c2_1_weight_class_matrix}, \ref{fig:c4_1_weight_class_matrix},3005and \ref{fig:c6_1_weight_class_matrix} continues with Figure~\ref{fig:c8_1_weight_class_matrix}.30063007\begin{figure}[!bhpt] % [H]3008\centering3009\begin{minipage}{.48\textwidth}3010\centering3011\includegraphics[width=.9\linewidth]{c8_1_weight_class_matrix.png}3012\captionof{figure}{$[f_{8,1}]$: weight classes. ~~~~\\~~~~}3013\label{fig:c8_1_weight_class_matrix}3014\end{minipage}%3015~~~~3016\begin{minipage}{.48\textwidth}3017\centering3018\includegraphics[width=.9\linewidth]{c8_1_bent_cayley_graph_index_matrix.png}3019\captionof{figure}{$[f_{8,1}]$: extended Cayley classes.}3020\label{fig:c8_1_bent_cayley_graph_index_matrix}3021\end{minipage}3022\end{figure}30233024% \end{frame}3025% \begin{frame}3026%3027\paragraph*{ET class $[f_{8,2}]$.}3028%3029This is the ET class of the bent function3030\small{}3031\begin{align*}3032f_{ 8 , 2 } &=3033\begin{array}{l}3034x_{0} x_{1} x_{2} + x_{0} x_{3} + x_{1} x_{4} + x_{2} x_{5} + x_{6} x_{7}.3035\end{array}3036\end{align*}3037\normalsize{}3038The ET class contains four extended Cayley classes as per Table~\ref{tab-c8_2_EC_classes}.30393040% f8_23041\begin{table}[!bhpt] % [H]3042% f_{8,2}:3043\small{}3044\begin{align*}3045\def\arraystretch{1.2}3046\begin{array}{|cccl|}3047\hline3048\text{Class} &3049\text{Parameters} &3050\text{2-rank} &3051\text{Clique polynomial}3052\\3053\hline30540 &3055(256, 120, 56, 56) &305610 &3057\begin{array}{l}3058245760t^{9} + 3317760t^{8} + 8847360t^{7}3059\,+3060\\306110321920t^{6} + 6193152t^{5} + 2007040t^{4}3062\,+3063\\3064286720t^{3} + 15360t^{2} + 256t + 13065\end{array}3066\\30671 &3068(256, 120, 56, 56) &306910 &3070\begin{array}{l}307149152t^{9} + 663552t^{8} + 2555904t^{7}3072\,+3073\\30745079040t^{6} + 4620288t^{5} + 1875968t^{4}3075\,+3076\\3077286720t^{3} + 15360t^{2} + 256t + 13078\end{array}3079\\30802 &3081(256, 136, 72, 72) &308210 &3083\begin{array}{l}3084327680t^{9} + 4055040t^{8} + 13828096t^{7}3085\,+3086\\308722183936t^{6} + 14319616t^{5} + 3891200t^{4}3088\,+3089\\3090417792t^{3} + 17408t^{2} + 256t + 13091\end{array}3092\\30933 &3094(256, 136, 72, 72) &309510 &3096\begin{array}{l}3097417792t^{8} + 3342336t^{7} + 11698176t^{6}3098\,+3099\\310011698176t^{5} + 3760128t^{4} + 417792t^{3}3101\,+3102\\310317408t^{2} + 256t + 13104\end{array}3105\\3106\hline3107\end{array}3108\end{align*}3109\caption{$[f_{8,2}]$ extended Cayley classes.}3110\label{tab-c8_2_EC_classes}3111\end{table}3112%\newpage31133114\begin{figure}[!ht] % [H]3115\centering3116\begin{minipage}{.48\textwidth}3117\centering3118\includegraphics[width=.9\linewidth]{c8_2_weight_class_matrix.png}3119\captionof{figure}{$[f_{8,2}]$: weight classes. ~~~~\\~~~~}3120\label{fig:c8_2_weight_class_matrix}3121\end{minipage}%3122~~~~3123\begin{minipage}{.48\textwidth}3124\centering3125\includegraphics[width=.9\linewidth]{c8_2_bent_cayley_graph_index_matrix.png}3126\captionof{figure}{$[f_{8,2}]$: extended Cayley classes.}3127\label{fig:c8_2_bent_cayley_graph_index_matrix}3128\end{minipage}3129\end{figure}3130~3131% \end{frame}3132\newpage3133The Cayley graph for class 0 is isomorphic to graph 0 of ET class $[f_{8,1}]$,3134This reflects the fact that $f_{8,1} \equiv f_{8,2}$, even though these two functions are not3135EA equivalent.3136This is therefore an example of an isomorphism between Cayley graphs of bent functions on3137$\F_2^8$ that is not a linear function.31383139The four extended Cayley classes are distributed between the two weight classes,3140as shown in Figures~\ref{fig:c8_2_weight_class_matrix} and~\ref{fig:c8_2_bent_cayley_graph_index_matrix}.31413142%\newpage3143% \begin{frame}3144%3145\paragraph*{ET class $[f_{8,3}]$.}3146%3147This is the ET class of the bent function3148\small{}3149\begin{align*}3150f_{ 8 , 3 } &=3151\begin{array}{l}3152x_{0} x_{1} x_{2} + x_{0} x_{6} + x_{1} x_{3} x_{4} + x_{1} x_{5} + x_{2} x_{3} + x_{4} x_{7}.3153\end{array}3154\end{align*}3155\normalsize{}3156The ET class contains six extended Cayley classes as per Table~\ref{tab-c8_3_EC_classes}.31573158% f8_33159\begin{table}[!bhpt] % [H]3160%3161% f_{8,3}:3162\small{}3163\begin{align*}3164\def\arraystretch{1.2}3165\begin{array}{|cccl|}3166\hline3167\text{Class} &3168\text{Parameters} &3169\text{2-rank} &3170\text{Clique polynomial}3171\\3172\hline31730 &3174(256, 120, 56, 56) &317512 &3176\begin{array}{l}317781920t^{9} + 1368064t^{8} + 4653056t^{7}3178\,+3179\\31807176192t^{6} + 5406720t^{5} + 1941504t^{4}3181\,+3182\\3183286720t^{3} + 15360t^{2} + 256t + 13184\end{array}3185\\31861 &3187(256, 136, 72, 72) &318812 &3189\begin{array}{l}3190294912t^{9} + 6299648t^{8} + 21692416t^{7}3191\,+3192\\319327951104t^{6} + 15630336t^{5} + 3956736t^{4}3194\,+3195\\3196417792t^{3} + 17408t^{2} + 256t + 13197\end{array}3198\\31992 &3200(256, 120, 56, 56) &320112 &3202\begin{array}{l}320316384t^{9} + 221184t^{8} + 1277952t^{7}3204\,+3205\\32063768320t^{6} + 4227072t^{5} + 1843200t^{4}3207\,+3208\\3209286720t^{3} + 15360t^{2} + 256t + 13210\end{array}3211\\32123 &3213(256, 136, 72, 72) &321412 &3215\begin{array}{l}3216262144t^{9} + 4399104t^{8} + 16220160t^{7}3217\,+3218\\321924281088t^{6} + 14974976t^{5} + 3923968t^{4}3220\,+3221\\3222417792t^{3} + 17408t^{2} + 256t + 13223\end{array}3224\\32254 &3226(256, 120, 56, 56) &322712 &3228\begin{array}{l}322949152t^{9} + 729088t^{8} + 2686976t^{7}3230\,+3231\\32325079040t^{6} + 4620288t^{5} + 1875968t^{4}3233\,+3234\\3235286720t^{3} + 15360t^{2} + 256t + 13236\end{array}3237\\32385 &3239(256, 136, 72, 72) &324012 &3241\begin{array}{l}3242196608t^{9} + 3399680t^{8} + 13172736t^{7}3243\,+3244\\324521659648t^{6} + 14319616t^{5} + 3891200t^{4}3246\,+3247\\3248417792t^{3} + 17408t^{2} + 256t + 13249\end{array}3250\\3251\hline3252\end{array}3253\end{align*}3254%3255\caption{$[f_{8,3}]$ extended Cayley classes.}3256\label{tab-c8_3_EC_classes}3257\end{table}32583259The six extended Cayley classes are distributed between the two weight classes,3260as shown in Figures~\ref{fig:c8_3_weight_class_matrix} and~\ref{fig:c8_3_bent_cayley_graph_index_matrix}.32613262\begin{figure}[!bhpt] % [H]3263\centering3264\begin{minipage}{.48\textwidth}3265\centering3266\includegraphics[width=.9\linewidth]{c8_3_weight_class_matrix.png}3267\captionof{figure}{$[f_{8,3}]$: weight classes. ~~~~\\~~~~}3268\label{fig:c8_3_weight_class_matrix}3269\end{minipage}%3270~~~~3271\begin{minipage}{.48\textwidth}3272\centering3273\includegraphics[width=.9\linewidth]{c8_3_bent_cayley_graph_index_matrix.png}3274\captionof{figure}{$[f_{8,3}]$: extended Cayley classes.}3275\label{fig:c8_3_bent_cayley_graph_index_matrix}3276\end{minipage}3277\end{figure}3278~3279% \end{frame}3280\newpage3281% \begin{frame}3282%3283\paragraph*{ET class $[f_{8,4}]$.}3284This is the ET class of the bent function3285\small{}3286\begin{align*}3287f_{ 8 , 4 } &=3288\begin{array}{l}3289x_{0} x_{1} x_{2} + x_{0} x_{2} + x_{0} x_{4} + x_{1} x_{3} x_{4} + x_{1} x_{5} + x_{2} x_{3}\, +3290x_{6} x_{7}.3291\end{array}3292\end{align*}3293\normalsize{}3294The ET class contains six extended Cayley classes as per Table~\ref{tab-c8_4_EC_classes}.32953296The six extended Cayley classes are distributed between the two weight classes,3297as shown in Figures~\ref{fig:c8_4_weight_class_matrix} and~\ref{fig:c8_4_bent_cayley_graph_index_matrix}.32983299% f8_43300\begin{table}[!bhpt] % [H]3301%3302\small{}3303\begin{align*}3304\def\arraystretch{1.2}3305\begin{array}{|cccl|}3306\hline3307\text{Class} &3308\text{Parameters} &3309\text{2-rank} &3310\text{Clique polynomial}3311\\3312\hline33130 &3314(256, 120, 56, 56) &331514 &3316\begin{array}{l}331769632t^{9} + 1099776t^{8} + 3784704t^{7}3318\,+3319\\33206160384t^{6} + 5013504t^{5} + 1908736t^{4}3321\,+3322\\3323286720t^{3} + 15360t^{2} + 256t + 13324\end{array}3325\\33261 &3327(256, 136, 72, 72) &332814 &3329\begin{array}{l}3330225280t^{9} + 4319232t^{8} + 16203776t^{7}3331\,+3332\\333324313856t^{6} + 14974976t^{5} + 3923968t^{4}3334\,+3335\\3336417792t^{3} + 17408t^{2} + 256t + 13337\end{array}3338\\33392 &3340(256, 120, 56, 56) &334114 &3342\begin{array}{l}33431536t^{10} + 15360t^{9} + 209920t^{8}3344\,+3345\\33461280000t^{7} + 3751936t^{6} + 4227072t^{5}3347\,+3348\\33491843200t^{4} + 286720t^{3} + 15360t^{2} + 256t + 13350\end{array}3351\\33523 &3353(256, 136, 72, 72) &335414 &3355\begin{array}{l}33567680t^{10} + 230400t^{9} + 4228096t^{8}3357\,+3358\\335916058368t^{7} + 24166400t^{6} + 14974976t^{5}3360\,+3361\\33623923968t^{4} + 417792t^{3} + 17408t^{2} + 256t + 13363\end{array}3364\\33654 &3366(256, 136, 72, 72) &336714 &3368\begin{array}{l}3369110592t^{9} + 2344960t^{8} + 10305536t^{7}3370\,+3371\\337218939904t^{6} + 13664256t^{5} + 3858432t^{4}3373\,+3374\\3375417792t^{3} + 17408t^{2} + 256t + 13376\end{array}3377\\33785 &3379(256, 120, 56, 56) &338014 &3381\begin{array}{l}338220480t^{9} + 337920t^{8} + 1556480t^{7}3383\,+3384\\33853932160t^{6} + 4227072t^{5} + 1843200t^{4}3386\,+3387\\3388286720t^{3} + 15360t^{2} + 256t + 13389\end{array}3390\\3391\hline3392\end{array}3393\end{align*}3394\caption{$[f_{8,4]}$ extended Cayley classes.}3395\label{tab-c8_4_EC_classes}3396\end{table}33973398\begin{figure}[!bhpt] % [H]3399\centering3400\begin{minipage}{.48\textwidth}3401\centering3402\includegraphics[width=.9\linewidth]{c8_4_weight_class_matrix.png}3403\captionof{figure}{$[f_{8,4}]$: weight classes. ~~~~\\~~~~}3404\label{fig:c8_4_weight_class_matrix}3405\end{minipage}%3406~~~~3407\begin{minipage}{.48\textwidth}3408\centering3409\includegraphics[width=.9\linewidth]{c8_4_bent_cayley_graph_index_matrix.png}3410\captionof{figure}{$[f_{8,4}]$: extended Cayley classes.}3411\label{fig:c8_4_bent_cayley_graph_index_matrix}3412\end{minipage}3413\end{figure}3414~3415% \end{frame}3416%%\newpage3417% \begin{frame}3418\paragraph*{ET class $[f_{8,5}]$.}3419%3420This is the ET class of the bent function3421\small{}3422\begin{align*}3423f_{ 8 , 5 } &=3424\begin{array}{l}3425x_{0} x_{1} x_{2} + x_{0} x_{6} + x_{1} x_{3} x_{4} + x_{1} x_{4} + x_{1} x_{5} + x_{2} x_{3} x_{5}3426+ x_{2} x_{4} + x_{3} x_{7}.3427\end{array}3428\end{align*}3429\normalsize{}3430The ET class contains 9 extended Cayley classes as per3431%Tables~\ref{tab-c8_5_EC_classes} and~\ref{tab-c8_5_EC_classes_2}.3432Table~\ref{tab-c8_5_EC_classes}.34333434% f8_53435\begin{table}[!bhpt] % [H]3436%3437\small{}3438\begin{align*}3439\def\arraystretch{1.2}3440\begin{array}{|cccl|}3441\hline3442\text{Class} &3443\text{Parameters} &3444\text{2-rank} &3445\text{Clique polynomial}3446\\3447\hline34480 &3449(256, 120, 56, 56) &345014 &3451\begin{array}{l}345232768t^{9} + 731136t^{8} + 3096576t^{7}3453\,+3454\\34555767168t^{6} + 5013504t^{5} + 1908736t^{4}3456\,+3457\\3458286720t^{3} + 15360t^{2} + 256t + 13459\end{array}3460\\34611 &3462(256, 120, 56, 56) &346314 &3464\begin{array}{l}346528672t^{9} + 534528t^{8} + 2211840t^{7}3466\,+3467\\34684718592t^{6} + 4620288t^{5} + 1875968t^{4}3469\,+3470\\3471286720t^{3} + 15360t^{2} + 256t + 13472\end{array}3473\\34742 &3475(256, 136, 72, 72) &347614 &3477\begin{array}{l}3478159744t^{9} + 4753408t^{8} + 19021824t^{7}3479\,+3480\\348126804224t^{6} + 15630336t^{5} + 3956736t^{4}3482\,+3483\\3484417792t^{3} + 17408t^{2} + 256t + 13485\end{array}3486\\34873 &3488(256, 120, 56, 56) &348914 &3490\begin{array}{l}349124576t^{9} + 526336t^{8} + 2342912t^{7}3492\,+3493\\34944849664t^{6} + 4620288t^{5} + 1875968t^{4}3495\,+3496\\3497286720t^{3} + 15360t^{2} + 256t + 13498\end{array}3499\\35004 &3501(256, 136, 72, 72) &350214 &3503\begin{array}{l}350490112t^{9} + 2795520t^{8} + 12402688t^{7}3505\,+3506\\350721168128t^{6} + 14319616t^{5} + 3891200t^{4}3508\,+3509\\3510417792t^{3} + 17408t^{2} + 256t + 13511\end{array}3512\\3513% \hline3514% \end{array}3515% \end{align*}3516% \caption{$[f_{8,5}]$ extended Cayley classes (part 1).}3517% \label{tab-c8_5_EC_classes}3518% \end{table}3519% %\newpage3520% \begin{table}[!bhpt] % [H]3521% \small{}3522% \begin{align*}3523% \def\arraystretch{1.2}3524% \begin{array}{|cccl|}3525% \hline3526% \text{Class} &3527% \text{Parameters} &3528% \text{2-rank} &3529% \text{Clique polynomial}3530% \\3531% \hline35325 &3533(256, 120, 56, 56) &353414 &3535\begin{array}{l}353616384t^{9} + 284672t^{8} + 1392640t^{7}3537\,+3538\\35393735552t^{6} + 4227072t^{5} + 1843200t^{4}3540\,+3541\\3542286720t^{3} + 15360t^{2} + 256t + 13543\end{array}3544\\35456 &3546(256, 136, 72, 72) &354714 &3548\begin{array}{l}3549131072t^{9} + 3577856t^{8} + 15319040t^{7}3550\,+3551\\355223855104t^{6} + 14974976t^{5} + 3923968t^{4}3553\,+3554\\3555417792t^{3} + 17408t^{2} + 256t + 13556\end{array}3557\\35587 &3559(256, 120, 56, 56) &356014 &3561\begin{array}{l}35621536t^{10} + 19456t^{9} + 279552t^{8}3563\,+3564\\35651394688t^{7} + 3751936t^{6} + 4227072t^{5}3566\,+3567\\35681843200t^{4} + 286720t^{3} + 15360t^{2} + 256t + 13569\end{array}3570\\35718 &3572(256, 136, 72, 72) &357314 &3574\begin{array}{l}35755632t^{10} + 148480t^{9} + 3621888t^{8}3576\,+3577\\357815206400t^{7} + 23773184t^{6} + 14974976t^{5}3579\,+3580\\35813923968t^{4} + 417792t^{3} + 17408t^{2} + 256t + 13582\end{array}3583\\3584\hline3585\end{array}3586\end{align*}3587%\caption{$[f_{8,5}]$ extended Cayley classes (part 2).}3588\caption{$[f_{8,5}]$ extended Cayley classes.}3589%\label{tab-c8_5_EC_classes_2}3590\label{tab-c8_5_EC_classes}3591\end{table}3592\newpage3593The 9 extended Cayley classes are distributed between the two weight classes,3594as shown in Figures~\ref{fig:c8_5_weight_class_matrix} and~\ref{fig:c8_5_bent_cayley_graph_index_matrix}.35953596\begin{figure}[!bhpt] % [H]3597\centering3598\begin{minipage}{.48\textwidth}3599\centering3600\includegraphics[width=.9\linewidth]{c8_5_weight_class_matrix.png}3601\captionof{figure}{$[f_{8,5}]$: weight classes. ~~~~\\~~~~}3602\label{fig:c8_5_weight_class_matrix}3603\end{minipage}%3604~~~~3605\begin{minipage}{.48\textwidth}3606\centering3607\includegraphics[width=.9\linewidth]{c8_5_bent_cayley_graph_index_matrix.png}3608\captionof{figure}{$[f_{8,5}]$: extended Cayley classes.}3609\label{fig:c8_5_bent_cayley_graph_index_matrix}3610\end{minipage}3611\end{figure}3612~3613% \end{frame}3614%\newpage3615% \begin{frame}3616\paragraph*{ET class $[f_{8,6}]$.}3617%3618%3619This is the ET class of the bent function3620\small{}3621\begin{align*}3622f_{ 8 , 6 } &=3623\begin{array}{l}3624x_{0} x_{1} x_{2} + x_{0} x_{2} + x_{0} x_{3} + x_{1} x_{3} x_{4} + x_{1} x_{6} + x_{2} x_{3} x_{5}3625+ x_{2} x_{4} + x_{5} x_{7}.3626\end{array}3627\end{align*}3628\normalsize{}3629The ET class contains 9 extended Cayley classes.3630% as per3631%Tables~\ref{tab-c8_6_EC_classes} and~\ref{tab-c8_6_EC_classes_2}.36323633%Each such isomorphism between3634%the Cayley graph of a function in $[f_{8,5}]$ and3635%the Cayley graph of a function in $[f_{8,6}]$ is not a linear function on $\F_2^8$.36363637%3638% % f8_63639% \begin{table}[!bhpt] % [H]3640% %3641% \small{}3642% \begin{align*}3643% \def\arraystretch{1.2}3644% \begin{array}{|cccl|}3645% \hline3646% \text{Class} &3647% \text{Parameters} &3648% \text{2-rank} &3649% \text{Clique polynomial}3650% \\3651% \hline3652% 0 &3653% (256, 120, 56, 56) &3654% 14 &3655% \begin{array}{l}3656% 32768t^{9} + 731136t^{8} + 3096576t^{7}3657% \,+3658% \\3659% 5767168t^{6} + 5013504t^{5} + 1908736t^{4}3660% \,+3661% \\3662% 286720t^{3} + 15360t^{2} + 256t + 13663% \end{array}3664% \\3665% 1 &3666% (256, 120, 56, 56) &3667% 14 &3668% \begin{array}{l}3669% 28672t^{9} + 534528t^{8} + 2211840t^{7}3670% \,+3671% \\3672% 4718592t^{6} + 4620288t^{5} + 1875968t^{4}3673% \,+3674% \\3675% 286720t^{3} + 15360t^{2} + 256t + 13676% \end{array}3677% \\3678% 2 &3679% (256, 136, 72, 72) &3680% 14 &3681% \begin{array}{l}3682% 159744t^{9} + 4753408t^{8} + 19021824t^{7}3683% \,+3684% \\3685% 26804224t^{6} + 15630336t^{5} + 3956736t^{4}3686% \,+3687% \\3688% 417792t^{3} + 17408t^{2} + 256t + 13689% \end{array}3690% \\3691% 3 &3692% (256, 120, 56, 56) &3693% 14 &3694% \begin{array}{l}3695% 1536t^{10} + 19456t^{9} + 279552t^{8}3696% \,+3697% \\3698% 1394688t^{7} + 3751936t^{6} + 4227072t^{5}3699% \,+3700% \\3701% 1843200t^{4} + 286720t^{3} + 15360t^{2} + 256t + 13702% \end{array}3703% \\3704% 4 &3705% (256, 136, 72, 72) &3706% 14 &3707% \begin{array}{l}3708% 5632t^{10} + 148480t^{9} + 3621888t^{8}3709% \,+3710% \\3711% 15206400t^{7} + 23773184t^{6} + 14974976t^{5}3712% \,+3713% \\3714% 3923968t^{4} + 417792t^{3} + 17408t^{2} + 256t + 13715% \end{array}3716% \\3717% \hline3718% \end{array}3719% \end{align*}3720% \caption{$[f_{8,6}]$ extended Cayley classes (part 1).}3721% \label{tab-c8_6_EC_classes}3722% \end{table}3723% %\newpage3724% \begin{table}[!bhpt] % [H]3725% \small{}3726% \begin{align*}3727% \def\arraystretch{1.2}3728% \begin{array}{|cccl|}3729% \hline3730% \text{Class} &3731% \text{Parameters} &3732% \text{2-rank} &3733% \text{Clique polynomial}3734% \\3735% \hline3736% 5 &3737% (256, 120, 56, 56) &3738% 14 &3739% \begin{array}{l}3740% 24576t^{9} + 526336t^{8} + 2342912t^{7}3741% \,+3742% \\3743% 4849664t^{6} + 4620288t^{5} + 1875968t^{4}3744% \,+3745% \\3746% 286720t^{3} + 15360t^{2} + 256t + 13747% \end{array}3748% \\3749% 6 &3750% (256, 120, 56, 56) &3751% 14 &3752% \begin{array}{l}3753% 16384t^{9} + 284672t^{8} + 1392640t^{7}3754% \,+3755% \\3756% 3735552t^{6} + 4227072t^{5} + 1843200t^{4}3757% \,+3758% \\3759% 286720t^{3} + 15360t^{2} + 256t + 13760% \end{array}3761% \\3762% 7 &3763% (256, 136, 72, 72) &3764% 14 &3765% \begin{array}{l}3766% 131072t^{9} + 3577856t^{8} + 15319040t^{7}3767% \,+3768% \\3769% 23855104t^{6} + 14974976t^{5} + 3923968t^{4}3770% \,+3771% \\3772% 417792t^{3} + 17408t^{2} + 256t + 13773% \end{array}3774% \\3775% 8 &3776% (256, 136, 72, 72) &3777% 14 &3778% \begin{array}{l}3779% 90112t^{9} + 2795520t^{8} + 12402688t^{7}3780% \,+3781% \\3782% 21168128t^{6} + 14319616t^{5} + 3891200t^{4}3783% \,+3784% \\3785% 417792t^{3} + 17408t^{2} + 256t + 13786% \end{array}3787% \\3788% \hline3789% \end{array}3790% \end{align*}3791% %3792% \caption{$[f_{8,6}]$ extended Cayley classes (part 2).}3793% %\caption{$[f_{8,6}]$ extended Cayley classes.}3794% \label{tab-c8_6_EC_classes_2}3795% %\label{tab-c8_6_EC_classes}3796% \end{table}37973798\begin{figure}[!hb] % [H]3799\centering3800\begin{minipage}{.48\textwidth}3801\centering3802\includegraphics[width=.9\linewidth]{c8_6_weight_class_matrix.png}3803\captionof{figure}{$[f_{8,6}]$: weight classes. ~~~~\\~~~~}3804\label{fig:c8_6_weight_class_matrix}3805\end{minipage}%3806~~~~3807\begin{minipage}{.48\textwidth}3808\centering3809\includegraphics[width=.9\linewidth]{c8_6_bent_cayley_graph_index_matrix.png}3810\captionof{figure}{$[f_{8,6}]$: extended Cayley classes.}3811\label{fig:c8_6_bent_cayley_graph_index_matrix}3812\end{minipage}3813\end{figure}3814% \end{frame}38153816The 9 extended Cayley classes are distributed between the two weight classes,3817as shown in Figures~\ref{fig:c8_6_weight_class_matrix} and~\ref{fig:c8_6_bent_cayley_graph_index_matrix}.38183819The 9 Cayley graphs corresponding to the 9 classes are isomorphic to those for the extended Cayley classes for $[f_{8,5}]$.3820The corresponding Cayley classes have the same frequency within each of these two ET classes.3821This correspondence is shown in Table~\ref{tab-c8_5-c8_6_EC_classes}.38223823Figures~\ref{fig:re8_5_bent_cayley_graph_index_matrix} and~\ref{fig:re8_5_bent_cayley_graph_index_matrix}3824show the 9 extended Cayley classes of each of $[f_{8,5}]$ and $[f_{8,6}]$ with corresponding Cayley classes3825given the same colour.38263827\begin{table}[!bhpt] % [H]3828%3829\small{}3830\begin{align*}3831\def\arraystretch{1.2}3832\begin{array}{|ccr|}3833\hline3834[f_{8,5}] &3835[f_{8,6}] &3836\text{Frequency}3837\\3838\hline38390 & 0 & 40963840\\38411 & 1 & 61443842\\38432 & 2 & 61443844\\38453 & 5 & 20483846\\38474 & 8 & 20483848\\38495 & 6 & 61443850\\38516 & 7 & 61443852\\38537 & 3 & 163843854\\38558 & 4 & 163843856\\3857\hline3858\end{array}3859\end{align*}3860\caption{Correspondence between $[f_{8,5}]$ and $[f_{8,6}]$ extended Cayley classes.}3861\label{tab-c8_5-c8_6_EC_classes}3862\end{table}38633864\begin{figure}[!bhpt] % [H]3865\centering3866\begin{minipage}{.48\textwidth}3867\centering3868\includegraphics[width=.9\linewidth]{re8_5_bent_cayley_graph_index_matrix.png}3869\captionof{figure}{$[f_{8,5}]$: extended Cayley classes (recoloured).}3870\label{fig:re8_5_bent_cayley_graph_index_matrix}3871\end{minipage}3872~~3873\begin{minipage}{.48\textwidth}3874\centering3875\includegraphics[width=.9\linewidth]{re8_6_bent_cayley_graph_index_matrix.png}3876\captionof{figure}{$[f_{8,6}]$: extended Cayley classes (recoloured).}3877\label{fig:re8_6_bent_cayley_graph_index_matrix}3878\end{minipage}3879\end{figure}3880\newpage3881The explanation for the correspondence between the Cayley classes of $f_{8,5}$ and $f_{8,6}$3882is quite simple. The functions $f_{8,5}$ and $f_{8,6}$ are EA equivalent,3883in fact general linear equivalent, and therefore Braeken's list of EA equivalence classes3884\cite[Section 5.5.2]{Bra06thesis} contains an error.38853886\begin{Theorem}3887\label{th-f8-5-f8-6-linearly-equiv}3888Functions $f_{8,5}$ and $f_{8,6}$ are general linear equivalent.3889\end{Theorem}38903891\begin{proof}38923893Apply the permutation $\pi := (x_0\ x_5\ x_4)(x_1\ x_2\ x_3)(x_6\ x_7)$ to3894\small{3895\begin{align*}3896f_{8,5}3897&=3898x_{0} x_{1} x_{2} + x_{0} x_{6} + x_{1} x_{3} x_{4} + x_{1} x_{4} + x_{1} x_{5} + x_{2} x_{3} x_{5} + x_{2} x_{4} + x_{3} x_{7}3899%\end{align*}3900%}\normalsize{}3901\intertext{\normalsize{to obtain}}3902%\small{3903%\begin{align*}3904\pi(f_{8,5})3905&=3906x_{5} x_{2} x_{3} + x_{5} x_{7} + x_{2} x_{1} x_{0} + x_{2} x_{0} + x_{2} x_{4} + x_{3} x_{1} x_{4} + x_{3} x_{0} + x_{1} x_{6}3907\\3908&=3909x_{0} x_{1} x_{2} + x_{0} x_{2} + x_{0} x_{3} + x_{1} x_{3} x_{4} + x_{1} x_{6} + x_{2} x_{3} x_{5} + x_{2} x_{4} + x_{5} x_{7}3910\\3911&= f_{8,6}.3912\end{align*}3913}\normalsize{}3914\end{proof}39153916%\newpage3917% \begin{frame}3918\paragraph*{ET class $[f_{8,7}]$.}3919%3920This is the ET class of the bent function3921\small{}3922\begin{align*}3923f_{ 8 , 7 } &=3924\begin{array}{l}3925x_{0} x_{1} x_{2} + x_{0} x_{1} + x_{0} x_{2} + x_{0} x_{3} + x_{1} x_{3} x_{4} + x_{1} x_{4}\, +3926x_{1} x_{5}\, +3927\\3928x_{2} x_{3} x_{5} + x_{2} x_{4} + x_{6} x_{7}.3929\end{array}3930\end{align*}3931\normalsize{}3932The ET class contains six extended Cayley classes as per Table~\ref{tab-c8_7_EC_classes}.39333934The six extended Cayley classes are distributed between the two weight classes,3935as shown in Figures~\ref{fig:c8_7_weight_class_matrix} and~\ref{fig:c8_7_bent_cayley_graph_index_matrix}.39363937% f8_73938\begin{table}[!bhpt] % [H]3939\small{}3940\begin{align*}3941\def\arraystretch{1.2}3942\begin{array}{|cccl|}3943\hline3944\text{Class} &3945\text{Parameters} &3946\text{2-rank} &3947\text{Clique polynomial}3948\\3949\hline39500 &3951(256, 120, 56, 56) &395216 &3953\begin{array}{l}395429696t^{9} + 655360t^{8} + 2789376t^{7}3955\,+3956\\39575332992t^{6} + 4816896t^{5} + 1892352t^{4}3958\,+3959\\3960286720t^{3} + 15360t^{2} + 256t + 13961\end{array}3962\\39631 &3964(256, 120, 56, 56) &396516 &3966\begin{array}{l}396720480t^{9} + 409600t^{8} + 1837056t^{7}3968\,+3969\\39704235264t^{6} + 4423680t^{5} + 1859584t^{4}3971\,+3972\\3973286720t^{3} + 15360t^{2} + 256t + 13974\end{array}3975\\39762 &3977(256, 136, 72, 72) &397816 &3979\begin{array}{l}3980143360t^{9} + 3981312t^{8} + 16697344t^{7}3981\,+3982\\398325108480t^{6} + 15302656t^{5} + 3940352t^{4}3984\,+3985\\3986417792t^{3} + 17408t^{2} + 256t + 13987\end{array}3988\\39893 &3990(256, 136, 72, 72) &399116 &3992\begin{array}{l}399364512t^{9} + 2316288t^{8} + 10932224t^{7}3994\,+3995\\399619783680t^{6} + 13991936t^{5} + 3874816t^{4}3997\,+3998\\3999417792t^{3} + 17408t^{2} + 256t + 14000\end{array}4001\\40024 &4003(256, 136, 72, 72) &400416 &4005\begin{array}{l}400692160t^{9} + 2979840t^{8} + 13608960t^{7}4007\,+4008\\400922388736t^{6} + 14647296t^{5} + 3907584t^{4}4010\,+4011\\4012417792t^{3} + 17408t^{2} + 256t + 14013\end{array}4014\\40155 &4016(256, 120, 56, 56) &401716 &4018\begin{array}{l}40196144t^{9} + 124928t^{8} + 944128t^{7}4020\,+4021\\40223219456t^{6} + 4030464t^{5} + 1826816t^{4}4023\,+4024\\4025286720t^{3} + 15360t^{2} + 256t + 14026\end{array}4027\\4028\hline4029\end{array}4030\end{align*}40314032\caption{$[f_{8,7}]$ extended Cayley classes.}4033\label{tab-c8_7_EC_classes}4034\end{table}40354036\begin{figure}[!bhpt] % [H]4037\centering4038\begin{minipage}{.48\textwidth}4039\centering4040\includegraphics[width=.9\linewidth]{c8_7_weight_class_matrix.png}4041\captionof{figure}{$[f_{8,7}]$: weight classes. ~~~~\\~~~~}4042\label{fig:c8_7_weight_class_matrix}4043\end{minipage}%4044~~~~4045\begin{minipage}{.48\textwidth}4046\centering4047\includegraphics[width=.9\linewidth]{c8_7_bent_cayley_graph_index_matrix.png}4048\captionof{figure}{$[f_{8,7}]$: extended Cayley classes.}4049\label{fig:c8_7_bent_cayley_graph_index_matrix}4050\end{minipage}4051\end{figure}40524053% \end{frame}4054\newpage4055% \begin{frame}4056\paragraph*{ET class $[f_{8,8}]$.}4057%4058This is the ET class of the bent function4059\small{}4060\begin{align*}4061f_{ 8 , 8 } &=4062\begin{array}{l}4063x_{0} x_{1} x_{2} + x_{0} x_{5} + x_{1} x_{3} x_{4} + x_{1} x_{6} + x_{2} x_{3} x_{5} + x_{2} x_{4}4064+ x_{3} x_{7}.4065\end{array}4066\end{align*}4067\normalsize{}4068The ET class contains six extended Cayley classes as per Table~\ref{tab-c8_8_EC_classes}.40694070The six extended Cayley classes are distributed between the two weight classes,4071as shown in Figures~\ref{fig:c8_8_weight_class_matrix} and~\ref{fig:c8_8_bent_cayley_graph_index_matrix}.40724073% f8_84074\begin{table}[!bhpt] % [H]4075\small{}4076\begin{align*}4077\def\arraystretch{1.2}4078\begin{array}{|cccl|}4079\hline4080\text{Class} &4081\text{Parameters} &4082\text{2-rank} &4083\text{Clique polynomial}4084\\4085\hline40860 &4087(256, 120, 56, 56) &408814 &4089\begin{array}{l}409032768t^{9} + 712704t^{8} + 3014656t^{7}4091\,+4092\\40935734400t^{6} + 5013504t^{5} + 1908736t^{4}4094\,+4095\\4096286720t^{3} + 15360t^{2} + 256t + 14097\end{array}4098\\40991 &4100(256, 120, 56, 56) &410114 &4102\begin{array}{l}410324576t^{9} + 466944t^{8} + 2064384t^{7}4104\,+4105\\41064685824t^{6} + 4620288t^{5} + 1875968t^{4}4107\,+4108\\4109286720t^{3} + 15360t^{2} + 256t + 14110\end{array}4111\\41122 &4113(256, 136, 72, 72) &411414 &4115\begin{array}{l}4116172032t^{9} + 5332992t^{8} + 20283392t^{7}4117\,+4118\\411927295744t^{6} + 15630336t^{5} + 3956736t^{4}4120\,+4121\\4122417792t^{3} + 17408t^{2} + 256t + 14123\end{array}4124\\41253 &4126(256, 136, 72, 72) &412714 &4128\begin{array}{l}4129147456t^{9} + 3858432t^{8} + 15990784t^{7}4130\,+4131\\413224150016t^{6} + 14974976t^{5} + 3923968t^{4}4133\,+4134\\4135417792t^{3} + 17408t^{2} + 256t + 14136\end{array}4137\\41384 &4139(256, 120, 56, 56) &414014 &4141\begin{array}{l}414216384t^{9} + 270336t^{8} + 1376256t^{7}4143\,+4144\\41453768320t^{6} + 4227072t^{5} + 1843200t^{4}4146\,+4147\\4148286720t^{3} + 15360t^{2} + 256t + 14149\end{array}4150\\41515 &4152(256, 136, 72, 72) &415314 &4154\begin{array}{l}4155163840t^{9} + 3858432t^{8} + 15532032t^{7}4156\,+4157\\415823887872t^{6} + 14974976t^{5} + 3923968t^{4}4159\,+4160\\4161417792t^{3} + 17408t^{2} + 256t + 14162\end{array}4163\\4164\hline4165\end{array}4166\end{align*}4167\caption{$[f_{8,8}]$ extended Cayley classes.}4168\label{tab-c8_8_EC_classes}4169\end{table}41704171\begin{figure}[!bhpt] % [H]4172\centering4173\begin{minipage}{.48\textwidth}4174\centering4175\includegraphics[width=.9\linewidth]{c8_8_weight_class_matrix.png}4176\captionof{figure}{$[f_{8,8}]$: weight classes. ~~~~\\~~~~}4177\label{fig:c8_8_weight_class_matrix}4178\end{minipage}%4179~~~~4180\begin{minipage}{.48\textwidth}4181\centering4182\includegraphics[width=.9\linewidth]{c8_8_bent_cayley_graph_index_matrix.png}4183\captionof{figure}{$[f_{8,8}]$: extended Cayley classes.}4184\label{fig:c8_8_bent_cayley_graph_index_matrix}4185\end{minipage}4186\end{figure}41874188% \end{frame}4189\newpage4190% \begin{frame}4191\paragraph*{ET class $[f_{8,9}]$.}4192%4193%4194This is the ET class of the bent function4195\small{}4196\begin{align*}4197f_{ 8 , 9 } &=4198\begin{array}{l}4199x_{0} x_{1} x_{6} + x_{0} x_{3} + x_{1} x_{4} + x_{2} x_{3} x_{6} + x_{2} x_{5} + x_{3} x_{4}\, +4200x_{4} x_{5} x_{6} + x_{6} x_{7}.4201\end{array}4202\end{align*}4203\normalsize{}4204The ET class contains 8 extended Cayley classes as per4205%Tables~\ref{tab-c8_9_EC_classes} and~\ref{tab-c8_9_EC_classes_2}.4206Table~\ref{tab-c8_9_EC_classes}.4207In 4 of these 8 classes, each bent function is extended Cayley equivalent to its dual.4208In the remaining 4 extended Cayley classes, the dual of every bent function in the class has a Cayley graph4209that is isomorphic to that of one other class. That is, the 4 Cayley classes form two duality pairs of classes.4210This correspondence between Cayley classes and Cayley classes of duals is shown in Table~\ref{tab-c8_9-dual-EC_classes}.42114212\begin{figure}[!bhpt] % [H]4213\centering4214\begin{minipage}{.48\textwidth}4215\centering4216\includegraphics[width=.9\linewidth]{c8_9_bent_cayley_graph_index_matrix.png}4217\captionof{figure}{$[f_{8,9}]$: 8 extended Cayley classes ~~ ~~~~ ~~~~ ~~~~~~~~~}4218\label{fig:c8_9_bent_cayley_graph_index_matrix}4219\end{minipage}4220~~4221\begin{minipage}{.48\textwidth}4222\centering4223\includegraphics[width=.9\linewidth]{c8_9_dual_cayley_graph_index_matrix.png}4224\captionof{figure}{$[f_{8,9}]$: 8 extended Cayley classes of dual bent functions}4225\label{fig:c8_9_dual_cayley_graph_index_matrix}4226\end{minipage}%4227\end{figure}42284229The 8 extended Cayley classes are distributed4230as shown in Figures~\ref{fig:c8_9_bent_cayley_graph_index_matrix} (classes) and~\ref{fig:c8_9_dual_cayley_graph_index_matrix}4231(classes of duals).42324233% f8_94234\begin{table}[!bhpt] % [H]4235\small{}4236\begin{align*}4237\def\arraystretch{1.2}4238\begin{array}{|cccl|}4239\hline4240\text{Class} &4241\text{Parameters} &4242\text{2-rank} &4243\text{Clique polynomial}4244\\4245\hline42460 &4247(256, 120, 56, 56) &424816 &4249\begin{array}{l}425045056t^{9} + 780288t^{8} + 2998272t^{7}4251\,+4252\\42535505024t^{6} + 4816896t^{5} + 1892352t^{4}4254\,+4255\\4256286720t^{3} + 15360t^{2} + 256t + 14257\end{array}4258\\42591 &4260(256, 120, 56, 56) &426116 &4262\begin{array}{l}426345056t^{9} + 780288t^{8} + 2998272t^{7}4264\,+4265\\42665505024t^{6} + 4816896t^{5} + 1892352t^{4}4267\,+4268\\4269286720t^{3} + 15360t^{2} + 256t + 14270\end{array}4271\\42722 &4273(256, 136, 72, 72) &427416 &4275\begin{array}{l}4276184320t^{9} + 3852288t^{8} + 14893056t^{7}4277\,+4278\\427923003136t^{6} + 14647296t^{5} + 3907584t^{4}4280\,+4281\\4282417792t^{3} + 17408t^{2} + 256t + 14283\end{array}4284\\42853 &4286(256, 136, 72, 72) &428716 &4288\begin{array}{l}4289184320t^{9} + 3852288t^{8} + 14893056t^{7}4290\,+4291\\429223003136t^{6} + 14647296t^{5} + 3907584t^{4}4293\,+4294\\4295417792t^{3} + 17408t^{2} + 256t + 14296\end{array}4297\\42984 &4299(256, 120, 56, 56) &430016 &4301\begin{array}{l}4302105984t^{8} + 976896t^{7} + 3440640t^{6}4303\,+4304\\43054128768t^{5} + 1835008t^{4} + 286720t^{3}4306\,+4307\\430815360t^{2} + 256t + 14309\end{array}4310\\4311% \hline4312% \end{array}4313% \end{align*}4314% \caption{$[f_{8,9}]$ extended Cayley classes (part 1).}4315% \label{tab-c8_9_EC_classes}4316% \end{table}4317% %%\newpage4318% \begin{table}[!bhpt] % [H]4319% \small{}4320% \begin{align*}4321% \def\arraystretch{1.2}4322% \begin{array}{|cccl|}4323% \hline4324% \text{Class} &4325% \text{Parameters} &4326% \text{2-rank} &4327% \text{Clique polynomial}4328% \\4329% \hline43305 &4331(256, 136, 72, 72) &433216 &4333\begin{array}{l}43349216t^{10} + 264192t^{9} + 4468224t^{8}4335\,+4336\\433716803840t^{7} + 24772608t^{6} + 15138816t^{5}4338\,+4339\\43403932160t^{4} + 417792t^{3} + 17408t^{2} + 256t + 14341\end{array}4342\\43436 &4344(256, 120, 56, 56) &434516 &4346\begin{array}{l}43479216t^{9} + 124416t^{8} + 976896t^{7}4348\,+4349\\43503440640t^{6} + 4128768t^{5} + 1835008t^{4}4351\,+4352\\4353286720t^{3} + 15360t^{2} + 256t + 14354\end{array}4355\\43567 &4357(256, 136, 72, 72) &435816 &4359\begin{array}{l}4360193536t^{9} + 4449792t^{8} + 16803840t^{7}4361\,+4362\\436324772608t^{6} + 15138816t^{5} + 3932160t^{4}4364\,+4365\\4366417792t^{3} + 17408t^{2} + 256t + 14367\end{array}4368\\4369\hline4370\end{array}4371\end{align*}4372%\caption{$[f_{8,9}]$ extended Cayley classes (part 2).}4373\caption{$[f_{8,9}]$ extended Cayley classes.}4374%\label{tab-c8_9_EC_classes_2}4375\label{tab-c8_9_EC_classes}4376\end{table}43774378\begin{table}4379\small{}4380\begin{align*}4381\def\arraystretch{1.2}4382\begin{array}{|ccc|}4383\hline4384[f_{8,9}] &4385[f_{8,9}] &4386\text{Frequency}4387\\4388&4389\text{duals} &4390\\4391\hline43920 & 1 & 92164393\\43941 & 0 & 92164395\\43962 & 3 & 71684397\\43983 & 2 & 71684399\\44004 & 4 & 81924401\\44025 & 5 & 81924403\\44046 & 6 & 81924405\\44067 & 7 & 81924407\\4408\hline4409\end{array}4410\end{align*}4411\caption{Correspondence between $[f_{8,9}]$ extended Cayley classes and $[f_{8,9}]$ dual extended Cayley classes.}4412\label{tab-c8_9-dual-EC_classes}4413\end{table}4414% \end{frame}4415\newpage4416% \begin{frame}4417\paragraph*{ET class $[f_{8,10}]$.}4418%4419This is the ET class of the bent function4420\small{}4421\begin{align*}4422f_{ 8 , 10 } :=4423\begin{array}{l}4424x_{0} x_{1} x_{2} + x_{0} x_{3} x_{6} + x_{0} x_{4} + x_{0} x_{5} + x_{1} x_{3} x_{4} + x_{1} x_{6}4425+ x_{2} x_{3} x_{5}\, +4426\\4427x_{2} x_{4} + x_{3} x_{7}.4428\end{array}4429\end{align*}4430\normalsize{}4431The ET class contains 10 extended Cayley classes as per4432Table~\ref{tab-c8_10_EC_classes}.4433%Tables~\ref{tab-c8_10_EC_classes} and~\ref{tab-c8_10_EC_classes_2}.4434In 4 of these 10 classes, each bent function is extended Cayley equivalent to its dual.4435In the remaining 6 extended Cayley classes, the dual of every bent function in the class has a Cayley graph4436that is isomorphic to that of one other class. That is, the 6 Cayley classes form three duality pairs of classes.4437This correspondence between Cayley classes and Cayley classes of duals is shown in Table~\ref{tab-c8_10-dual-EC_classes}.44384439The 10 extended Cayley classes are distributed4440as shown in Figures~\ref{fig:c8_10_bent_cayley_graph_index_matrix} (classes) and~\ref{fig:c8_10_dual_cayley_graph_index_matrix}4441(classes of duals).4442% f8_104443\begin{table}[!bhpt] % [H]4444\small{}4445\begin{align*}4446\def\arraystretch{1.2}4447\begin{array}{|cccl|}4448\hline4449\text{Class} &4450\text{Parameters} &4451\text{2-rank} &4452\text{Clique polynomial}4453\\4454\hline44550 &4456(256, 120, 56, 56) &445716 &4458\begin{array}{l}445916384t^{9} + 464896t^{8} + 2310144t^{7}4460\,+4461\\44625046272t^{6} + 4816896t^{5} + 1892352t^{4}4463\,+4464\\4465286720t^{3} + 15360t^{2} + 256t + 14466\end{array}4467\\44681 &4469(256, 120, 56, 56) &447016 &4471\begin{array}{l}447216384t^{9} + 464896t^{8} + 2310144t^{7}4473\,+4474\\44755046272t^{6} + 4816896t^{5} + 1892352t^{4}4476\,+4477\\4478286720t^{3} + 15360t^{2} + 256t + 14479\end{array}4480\\44812 &4482(256, 120, 56, 56) &448316 &4484\begin{array}{l}448512288t^{9} + 301056t^{8} + 1589248t^{7}4486\,+4487\\44884128768t^{6} + 4423680t^{5} + 1859584t^{4}4489\,+4490\\4491286720t^{3} + 15360t^{2} + 256t + 14492\end{array}4493\\44943 &4495(256, 120, 56, 56) &449616 &4497\begin{array}{l}449812288t^{9} + 301056t^{8} + 1589248t^{7}4499\,+4500\\45014128768t^{6} + 4423680t^{5} + 1859584t^{4}4502\,+4503\\4504286720t^{3} + 15360t^{2} + 256t + 14505\end{array}4506\\45074 &4508(256, 136, 72, 72) &450916 &4510\begin{array}{l}4511110592t^{9} + 4159488t^{8} + 17285120t^{7}4512\,+4513\\451425296896t^{6} + 15302656t^{5} + 3940352t^{4}4515\,+4516\\4517417792t^{3} + 17408t^{2} + 256t + 14518\end{array}4519\\4520% \hline4521% \end{array}4522% \end{align*}4523% \caption{$[f_{8,10}]$ extended Cayley classes (part 1).}4524% \label{tab-c8_10_EC_classes}4525% \end{table}4526% %%\newpage4527% \begin{table}[!bhpt] % [H]4528% \small{}4529% \begin{align*}4530% \def\arraystretch{1.2}4531% \begin{array}{|cccl|}4532% \hline4533% \text{Class} &4534% \text{Parameters} &4535% \text{2-rank} &4536% \text{Clique polynomial}4537% \\4538% \hline45395 &4540(256, 136, 72, 72) &454116 &4542\begin{array}{l}4543110592t^{9} + 4159488t^{8} + 17285120t^{7}4544\,+4545\\454625296896t^{6} + 15302656t^{5} + 3940352t^{4}4547\,+4548\\4549417792t^{3} + 17408t^{2} + 256t + 14550\end{array}4551\\45526 &4553(256, 120, 56, 56) &455416 &4555\begin{array}{l}45562048t^{9} + 167424t^{8} + 1091584t^{7}4557\,+4558\\45593440640t^{6} + 4128768t^{5} + 1835008t^{4}4560\,+4561\\4562286720t^{3} + 15360t^{2} + 256t + 14563\end{array}4564\\45657 &4566(256, 136, 72, 72) &456716 &4568\begin{array}{l}45697168t^{10} + 143360t^{9} + 3804672t^{8}4570\,+4571\\457215886336t^{7} + 24313856t^{6} + 15138816t^{5}4573\,+4574\\45753932160t^{4} + 417792t^{3} + 17408t^{2} + 256t + 14576\end{array}4577\\45788 &4579(256, 120, 56, 56) &458016 &4581\begin{array}{l}45829216t^{9} + 181760t^{8} + 1091584t^{7}4583\,+4584\\45853440640t^{6} + 4128768t^{5} + 1835008t^{4}4586\,+4587\\4588286720t^{3} + 15360t^{2} + 256t + 14589\end{array}4590\\45919 &4592(256, 136, 72, 72) &459316 &4594\begin{array}{l}4595107520t^{9} + 3790336t^{8} + 15886336t^{7}4596\,+4597\\459824313856t^{6} + 15138816t^{5} + 3932160t^{4}4599\,+4600\\4601417792t^{3} + 17408t^{2} + 256t + 14602\end{array}4603\\4604\hline4605\end{array}4606\end{align*}4607%\caption{$[f_{8,10}]$ extended Cayley classes (part 2).}4608\caption{$[f_{8,10}]$ extended Cayley classes.}4609%\label{tab-c8_10_EC_classes_2}4610\label{tab-c8_10_EC_classes}4611\end{table}46124613\begin{table}4614\small{}4615\begin{align*}4616\def\arraystretch{1.2}4617\begin{array}{|ccc|}4618\hline4619[f_{8,10}] &4620[f_{8,10}] &4621\text{Frequency}4622\\4623&4624\text{duals} &4625\\4626\hline46270 & 1 & 20484628\\46291 & 0 & 20484630\\46312 & 3 & 71684632\\46333 & 2 & 71684634\\46354 & 5 & 71684636\\46375 & 4 & 71684638\\46396 & 6 & 81924640\\46417 & 7 & 81924642\\46438 & 8 & 81924644\\46459 & 9 & 81924646\\4647\hline4648\end{array}4649\end{align*}4650\caption{Correspondence between $[f_{8,10}]$ extended Cayley classes and $[f_{8,10}]$ dual extended Cayley classes.}4651\label{tab-c8_10-dual-EC_classes}4652\end{table}46534654\begin{figure}[!bhpt] % [H]4655\centering4656\begin{minipage}{.48\textwidth}4657\centering4658\includegraphics[width=.9\linewidth]{c8_10_bent_cayley_graph_index_matrix.png}4659\captionof{figure}{$[f_{8,10}]$: extended Cayley classes.}4660\label{fig:c8_10_bent_cayley_graph_index_matrix}4661\end{minipage}4662~~4663\begin{minipage}{.48\textwidth}4664\centering4665\includegraphics[width=.9\linewidth]{c8_10_dual_cayley_graph_index_matrix.png}4666\captionof{figure}{$[f_{8,10}]$: extended Cayley classes of dual bent functions.}4667\label{fig:c8_10_dual_cayley_graph_index_matrix}4668\end{minipage}%4669\end{figure}4670% \end{frame}4671% \end{colortheme}4672% \begin{colortheme}{seagull}4673% \begin{frame}4674\newpage46754676\subsection{Two sequences of bent functions}46774678As stated in the introduction, in a recent paper \cite{Leo17Hurwitz},4679the author found an example of two infinite series of bent functions whose4680Cayley graphs have the same strongly regular parameters at each dimension,4681but are not isomorphic if the dimension is 8 or more.4682The sequences are $\sigma_m$ and $\tau_m$ for $m \geqslant 1$,4683whose definitions are reproduced here from \cite{Leo15Twin}.46844685\begin{definition}4686\label{def-sigma-tau}4687The sign-of-square function $\sigma_m : \Z_2^{2 m} \To \Z_2$ is defined as follows:46884689For $i \in \Z_{2^{2m}},$ $\sigma_m(i) = 1$ if and only if the number of46901 digits in the base 4 representation of $i$ is odd.46914692The non-diagonal-symmetry function $\tau_m \colon \Z_{2^{2 m}} \To \Z_2$4693is defined as follows.46944695For $i$ in $\Z_2^2$:4696\begin{align*}4697&\tau_1(i) :=4698\begin{cases}46991 &\text{if~}i = 10,4700\\47010 &\text{otherwise}.4702\end{cases}4703\end{align*}47044705For $i$ in $\Z_2^{2 m - 2}$:4706\begin{align*}4707&\tau_m (00 \odot i) := \tau_{m-1}(i),4708\\4709&\tau_m (01 \odot i) := \sigma_{m-1}(i),4710\\4711&\tau_m (10 \odot i) := \sigma_{m-1}(i) + 1,4712\\4713&\tau_m (11 \odot i) := \tau_{m-1}(i).4714\end{align*}4715where $\odot$ denotes concatenation of bit vectors, and $\sigma$ is the sign-of-square function, as above.4716\end{definition}47174718As shown in \cite{Leo15Twin}, both sequences produce Cayley graphs whose strongly regular parameters are4719\begin{align*}4720(v_m,k_m,\lambda_m=\mu_m) &= (4^m, 2^{2 m - 1} - 2^{m-1}, 2^{2 m - 2} - 2^{m-1}).4721\end{align*}47224723%\begin{colortheme}{jubata}47244725%\begin{frame}4726%\frametitle{For 2 dimensions: $[\sigma_1]$ and $[\tau_1]$}4727\begin{figure}[!ht]4728\centering4729\begin{minipage}{.48\textwidth}4730\centering4731\includegraphics[width=.9\linewidth]{sigma_1_bent_cayley_graph_index_matrix.png}4732\captionof{figure}{$[\sigma_1]$:\\2 extended Cayley classes}4733\label{fig:sigma_1_bent_cayley_graph_index_matrix}4734\end{minipage}%4735\begin{minipage}{.48\textwidth}4736\centering4737\includegraphics[width=.9\linewidth]{tau_1_bent_cayley_graph_index_matrix.png}4738\captionof{figure}{$[\tau_1]$:\\2 extended Cayley classes}4739\label{fig:tau_1_bent_cayley_graph_index_matrix}4740\end{minipage}4741\end{figure}4742%\end{frame}47434744%\begin{frame}4745%\frametitle{For 4 dimensions: $[\sigma_2]$ and $[\tau_2]$}4746\begin{figure}[!hb]4747\centering4748\begin{minipage}{.48\textwidth}4749\centering4750\includegraphics[width=.9\linewidth]{sigma_2_bent_cayley_graph_index_matrix.png}4751\captionof{figure}{$[\sigma_2]$:\\2 extended Cayley classes}4752\label{fig:sigma_2_bent_cayley_graph_index_matrix}4753\end{minipage}%4754\begin{minipage}{.48\textwidth}4755\centering4756\includegraphics[width=.9\linewidth]{tau_2_bent_cayley_graph_index_matrix.png}4757\captionof{figure}{$[\tau_2]$:\\2 extended Cayley classes}4758\label{fig:tau_2_bent_cayley_graph_index_matrix}4759\end{minipage}4760\end{figure}4761%\end{frame}47624763Figures \ref{fig:sigma_1_bent_cayley_graph_index_matrix} to \ref{fig:tau_4_bent_cayley_graph_index_matrix}4764illustrate the extended Cayley classes within the ET classes of each of function $\sigma_m$ and $\tau_m$ for $m$ from 1 to 4.4765Note that $\tau_3$ has degree 3, and $\tau_4$ has degree 4.4766As shown in \cite{Leo17Hurwitz},4767The functions $\sigma_m$ and $\tau_m$ are extended Cayley equivalent for $m$ from 1 to 3, but are inequivalent for $m>3$.47684769%\begin{frame}4770%\frametitle{For 6 dimensions: $[\sigma_3]$ and $[\tau_3]$}4771\begin{figure}[!ht]4772\centering4773\begin{minipage}{.48\textwidth}4774\centering4775\includegraphics[width=.9\linewidth]{sigma_3_bent_cayley_graph_index_matrix.png}4776\captionof{figure}{$[\sigma_3]$:\\2 extended Cayley classes}4777\label{fig:sigma_3_bent_cayley_graph_index_matrix}4778\end{minipage}%4779\begin{minipage}{.48\textwidth}4780\centering4781\includegraphics[width=.9\linewidth]{tau_3_bent_cayley_graph_index_matrix.png}4782\captionof{figure}{$[\tau_3]$:\\3 extended Cayley classes}4783\label{fig:tau_3_bent_cayley_graph_index_matrix}4784\end{minipage}4785\end{figure}4786%\end{frame}47874788%\begin{frame}4789%\frametitle{For 8 dimensions: $[\sigma_4]$ and $[\tau_4]$}4790\begin{figure}[!hb]4791\centering4792\begin{minipage}{.48\textwidth}4793\centering4794\includegraphics[width=.9\linewidth]{sigma_4_bent_cayley_graph_index_matrix.png}4795\captionof{figure}{$[\sigma_4]$:\\2 extended Cayley classes}4796\label{fig:sigma_4_bent_cayley_graph_index_matrix}4797\end{minipage}%4798\begin{minipage}{.48\textwidth}4799\centering4800\includegraphics[width=.9\linewidth]{tau_4_bent_cayley_graph_index_matrix.png}4801\captionof{figure}{$[\tau_4]$:\\5 extended Cayley classes}4802\label{fig:tau_4_bent_cayley_graph_index_matrix}4803\end{minipage}4804\end{figure}4805%\end{frame}4806%\end{colortheme}4807~4808\newpage4809~4810\newpage4811~4812\newpage4813%\paragraph*{Example CAST-128 S-Box ET class $[cast128_{1,0}]$.}4814\subsection{CAST-128 S-boxes}4815%4816The CAST-128 encryption algorithm is used in PGP and elsewhere \cite{Ada97}.4817The algorithm uses 8 S-boxes, each of which consists of 32 Boolean bent functions in 8 dimensions,4818with degree 4, making 256 bent functions in total.4819The full CAST-128 algorithm, including the contents of the S-boxes,4820is published as IETF Request For Comments 2144 \cite{RFC2144}.48214822The bent function $cast128_{1,0}$ is the first bent function of S-box number 1 of CAST-128.4823Its definition by algebraic normal form is4824\footnotesize{}4825\begin{align*}4826cast128_{1,0} :=4827\begin{array}{l}4828x_{0} x_{1} x_{2} x_{3} + x_{0} x_{1} x_{2} x_{4} + x_{0} x_{1} x_{2} x_{5} + x_{0} x_{1} x_{2} + x_{0} x_{1} x_{3} x_{5} + x_{0} x_{1} x_{3} x_{6}\, +4829\\4830x_{0} x_{1} x_{3} + x_{0} x_{1} x_{5} x_{6} + x_{0} x_{1} x_{6} + x_{0} x_{1} x_{7} + x_{0} x_{2} x_{3} x_{4} + x_{0} x_{2} x_{3}\, +4831\\4832x_{0} x_{2} x_{4} x_{5} + x_{0} x_{2} x_{5} x_{7} + x_{0} x_{2} x_{6} + x_{0} x_{2} x_{7} + x_{0} x_{2} + x_{0} x_{3} x_{4} x_{5}\, +4833\\4834x_{0} x_{3} x_{4} x_{6} + x_{0} x_{3} x_{5} x_{6} + x_{0} x_{3} + x_{0} x_{4} x_{5} x_{6} + x_{0} x_{4} x_{5} x_{7} + x_{0} x_{4} x_{5}\, +4835\\4836x_{0} x_{4} x_{6} + x_{0} x_{4} + x_{0} x_{5} x_{6} + x_{0} x_{5} x_{7} + x_{0} x_{6} + x_{0} x_{7} + x_{0} + x_{1} x_{2} x_{4} x_{6}\, +4837\\4838x_{1} x_{2} x_{4} x_{7} + x_{1} x_{2} x_{5} x_{6} + x_{1} x_{2} x_{7} + x_{1} x_{3} x_{4} x_{6} + x_{1} x_{3} x_{4} x_{7} + x_{1} x_{3} x_{5}\, +4839\\4840x_{1} x_{3} x_{7} + x_{1} x_{4} x_{5} x_{7} + x_{1} x_{4} x_{6} + x_{1} x_{5} x_{6} + x_{1} + x_{2} x_{3} x_{4} x_{6} + x_{2} x_{3} x_{4}\, +4841\\4842x_{2} x_{3} x_{5} x_{6} + x_{2} x_{3} x_{5} + x_{2} x_{3} x_{7} + x_{2} x_{4} + x_{2} x_{5} x_{6} + x_{2} x_{5} + x_{2} + x_{3} x_{4} x_{5} x_{6}\, +4843\\4844x_{3} x_{5} x_{6} + x_{3} x_{5} x_{7} + x_{3} x_{6} + x_{3} + x_{4} + x_{6} x_{7}.4845\end{array}4846\end{align*}4847\normalsize{}48484849This bent function $cast128_{1,0}$ is prolific:4850the ET class $[cast128_{1,0}]$ contains the maximum possible number of Cayley classes,4851that is $65\,536$.4852The duals of the bent functions $[cast128_{1,0}]$ give another $65\,536$ extended Cayley classes.4853In other words, no bent function in $[cast128_{1,0}]$ is EA equivalent to its dual.4854The two weight classes of $[cast128_{1,0}]$ are4855shown in Figure~\ref{fig:cast128_1_0_weight_class_matrix}.48564857\begin{figure}[!hb]4858\centering4859\begin{minipage}{.48\textwidth}4860\centering4861\includegraphics[width=.9\linewidth]{cast128_1_0_weight_class_matrix.png}4862\captionof{figure}{$[cast128_{1,0}]$:\\Weight classes ~~~~~~ ~~~~~~~~\\~\\Colormap: gist\_stern.}4863\label{fig:cast128_1_0_weight_class_matrix}4864\end{minipage}4865\begin{minipage}{.48\textwidth}4866\centering4867\includegraphics[width=.9\linewidth]{cast128_1_0_bent_cayley_graph_index_matrix.png}4868\captionof{figure}{$[cast128_{1,0}]$:\\$65\,536$ extended Cayley classes.\\Total including duals is $131\,072$.4869\\Colormap: jet.}4870\label{fig:cast128_1_0_bent_cayley_graph_index_matrix}4871\end{minipage}%4872\end{figure}4873%\frametitle{CAST-128 ET classes $[cast128_{2,1}]$ and $[cast128_{2,16}]$}48744875\begin{figure}[!ht]4876\centering4877\begin{minipage}{.48\textwidth}4878\centering4879\includegraphics[width=.9\linewidth]{cast128_2_1_bent_cayley_graph_index_matrix.png}4880\captionof{figure}{$[cast128_{2,1}]$:\\$8\,256$ extended Cayley classes.\\Total including duals is $8\,256$.4881\\Colormap: jet.}4882\label{fig:cast128_2_1_bent_cayley_graph_index_matrix}4883\end{minipage}%4884\begin{minipage}{.48\textwidth}4885\centering4886\includegraphics[width=.9\linewidth]{cast128_2_16_bent_cayley_graph_index_matrix.png}4887\captionof{figure}{$[cast128_{2,16}]$:\\$32\,768$ extended Cayley classes.\\Total including duals is $65\,536$.4888\\Colormap: jet.}4889\label{fig:cast128_2_16_bent_cayley_graph_index_matrix}4890\end{minipage}%4891\end{figure}4892%\frametitle{CAST-128 ET classes $[cast128_{4,27}]$ and $[cast128_{5,16}]$}4893\begin{figure}[!hb]4894\centering4895\begin{minipage}{.48\textwidth}4896\centering4897\includegraphics[width=.9\linewidth]{cast128_4_27_bent_cayley_graph_index_matrix.png}4898\captionof{figure}{$[cast128_{4,27}]$:\\$65\,536$ extended Cayley classes.\\Total including duals is $65\,536$.4899\\Colormap: jet.}4900\label{fig:cast128_4_27_bent_cayley_graph_index_matrix}4901\end{minipage}%4902\begin{minipage}{.48\textwidth}4903\centering4904\includegraphics[width=.9\linewidth]{cast128_5_16_bent_cayley_graph_index_matrix.png}4905\captionof{figure}{$[cast128_{5,16}]$:\\$33\,280$ extended Cayley classes.\\Total including duals is $66\,560$.4906\\Colormap: jet.}4907\label{fig:cast128_5_16_bent_cayley_graph_index_matrix}4908\end{minipage}%4909\end{figure}4910%\frametitle{CAST-128 ET classes $[cast128_{5,27}]$ and $[cast128_{6,17}]$}4911\begin{figure}[!ht]4912\centering4913\begin{minipage}{.48\textwidth}4914\centering4915\includegraphics[width=.9\linewidth]{cast128_5_27_bent_cayley_graph_index_matrix.png}4916\captionof{figure}{$[cast128_{5,27}]$:\\$6\,144$ extended Cayley classes.\\Total including duals is $6\,144$.4917\\Colormap: jet.}4918\label{fig:cast128_5_27_bent_cayley_graph_index_matrix}4919\end{minipage}%4920\begin{minipage}{.48\textwidth}4921\centering4922\includegraphics[width=.9\linewidth]{cast128_6_17_bent_cayley_graph_index_matrix.png}4923\captionof{figure}{$[cast128_{6,17}]$:\\$65\,536$ extended Cayley classes.\\Total including duals is $65\,536$.4924\\Colormap: jet.}4925\label{fig:cast128_6_17_bent_cayley_graph_index_matrix}4926\end{minipage}%4927\end{figure}4928%\frametitle{CAST-128 ET classes $[cast128_{7,15}]$ and $[cast128_{7,21}]$}4929\begin{figure}[!ht]4930\centering4931\begin{minipage}{.48\textwidth}4932\centering4933\includegraphics[width=.9\linewidth]{cast128_7_15_bent_cayley_graph_index_matrix.png}4934\captionof{figure}{$[cast128_{7,15}]$:\\$32\,768$ extended Cayley classes.\\Total including duals is $65\,536$.4935\\Colormap: jet.}4936\label{fig:cast128_7_15_bent_cayley_graph_index_matrix}4937\end{minipage}%4938\begin{minipage}{.48\textwidth}4939\centering4940\includegraphics[width=.9\linewidth]{cast128_7_21_bent_cayley_graph_index_matrix.png}4941\captionof{figure}{$[cast128_{7,21}]$:\\$32\,768$ extended Cayley classes.\\Total including duals is $65\,536$.4942\\Colormap: jet.}4943\label{fig:cast128_7_21_bent_cayley_graph_index_matrix}4944\end{minipage}%4945\end{figure}49464947Of the 256 bent functions that make up the 8 S-boxes of CAST-128, 248 are like $cast128_{1,0}$,4948they are prolific and are not EA equivalent to their duals.4949The remaining 8 bent functions are exceptional.4950Both $cast128_{4,27}$ and $cast128_{6,17}$ are prolific, but both are EA equivalent to their duals.4951\newpage49524953The other 6 bent functions are not prolific and the number of Cayley classes for each is given in the captions to4954Figures \ref{fig:cast128_2_1_bent_cayley_graph_index_matrix} to \ref{fig:cast128_7_21_bent_cayley_graph_index_matrix}.4955\subsection{Partial spread bent functions.}49564957According to Langevin and Hou \cite{LanH11counting}4958there are $70\,576\,747\,237\,594\,114\,392\,064 \approx 2^{75.9}$ \Emph{partial spread} bent functions in4959dimension 8,4960contained in $14\,758$ EA classes.4961The EA class representatives are listed at Langevin's web page \cite{Lan10psf}.4962On this web page,4963the file \texttt{psf-8.txt} contains details of $9316$ representatives that are4964$\mathcal{PS}^{(-)}$ bent functions, all of degree 4; and4965the file \texttt{psf-9.txt} contains details of $5442$ representatives that are4966$\mathcal{PS}^{(+)}$ bent functions, of which $5440$ are of degree 4.49674968Preliminary calculations using SageMath on the NCI Raijin supercomputer indicate that4969\begin{enumerate}4970\item4971The $5442$ EA classes of $\mathcal{PS}^{(+)}$ bent functions of dimension 84972contain $296\,594\,720$ extended Cayley classes, assuming that each extended Cayley class appears in only one of the EA classes.4973\item4974If the duals of the $5442$ representatives, and their corresponding EA classes are included,4975the total number of extended Cayley classes is $541\,700\,450$, under the same assumption.4976\item4977Of the $5442$ representatives,4978$3434$ are prolific and not EA equivalent to their dual,4979$582$ are prolific and are EA equivalent to their dual,4980and the EA classes of the remaining $1426$ each contain less than $65\,536$ extended Cayley classes.4981\end{enumerate}4982The classifications of these $5442$ EA classes take up 2.1TB of space on the NCI Raijin supercomputer,4983and it is intended that these classifications will be incorporated into a public database, if support can be found to maintain it4984\cite{Leo18Database}.4985%\slidecite{Langevin and Hou 2011}4986% \end{frame}4987% \end{colortheme}4988% \begin{colortheme}{jubata}4989% \begin{frame}4990%\paragraph*{Example partial spread ET class $[psf_{9,5439}]$.}49914992%There are obviously too many bent functions of these two types to list here, but there is room for one example.49934994One example classification from $\mathcal{PS}^{(+)}$ in dimension 8 is that of $psf_{9,5439}$.4995The bent function $psf_{9,5439}$ is listed as function number $5439$ in \texttt{psf-9.txt},4996and is a $\mathcal{PS}^{(+)}$ bent function of degree 4.4997Its ET class $[psf_{9,5439}]$ contains 16 extended Cayley classes,4998of which 6 form three duality pairs similar to those seen in $[f_{8,9}]$ and $[f_{8,10}]$ above.49995000The 16 extended Cayley classes are distributed5001as shown in Figures~\ref{fig:psf_9_5439_bent_cayley_graph_index_matrix} (classes) and~\ref{fig:psf_9_5439_dual_cayley_graph_index_matrix}5002(classes of duals).50035004\begin{figure}[!ht] % [H]5005\centering5006\begin{minipage}{.48\textwidth}5007\centering5008\includegraphics[width=.9\linewidth]{psf_9_5439_bent_cayley_graph_index_matrix.png}5009\captionof{figure}{$[psf_{9,5439}]$:\\extended Cayley classes ~~ ~~~~ ~~~~\\~~~~~~~~~}5010\label{fig:psf_9_5439_bent_cayley_graph_index_matrix}5011\end{minipage}5012~~5013\begin{minipage}{.48\textwidth}5014\centering5015\includegraphics[width=.9\linewidth]{psf_9_5439_dual_cayley_graph_index_matrix.png}5016\captionof{figure}{$[psf_{9,5439}]$:\\extended Cayley classes of dual bent\\functions}5017\label{fig:psf_9_5439_dual_cayley_graph_index_matrix}5018\end{minipage}%5019\end{figure}50205021% \end{frame}50225023%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%50245025%\addtocontents{toc}{\vspace{0.5cm}}50265027\newpage50285029\bibliographystyle{abbrv-par}50305031\bibliography{bib}50325033%\input{Leopardi-Bent-functions-Cayley-graphs.bbl}503450355036