"Talk on Siegel theta series and modular forms"의 두 판 사이의 차이

수학노트
둘러보기로 가기 검색하러 가기
imported>Pythagoras0
 
(다른 사용자 한 명의 중간 판 188개는 보이지 않습니다)
1번째 줄: 1번째 줄:
 
==overview==
 
==overview==
* Siegel theta series
+
* Siegel theta series  
 
* Siegel modular forms
 
* Siegel modular forms
 
* Siegel-Weil formula
 
* Siegel-Weil formula
* {{수학노트|url=지겔_모듈라_형식}}
 
* {{수학노트|url=격자의_지겔_세타_급수}}
 
* {{수학노트|url=리만_곡면의_주기_행렬과_겹선형_관계_(bilinear_relation)}}
 
  
 +
==modular forms==
 +
* <math>\mathbb{H}=\{\tau\in \mathbb{C}|\Im \tau>0\}</math>
 +
* modular group <math>\Gamma=SL(2, \mathbb Z) = \left \{ \left. \left ( \begin{array}{cc}a & b \\ c & d \end{array} \right )\right| a, b, c, d \in \mathbb Z,\ ad-bc = 1 \right \}</math>
 +
* <math>\operatorname{PSL}(2,\mathbb{Z})=\operatorname{SL}(2,\mathbb{Z})/\{\pm I\}</math> acts on <math>\mathbb{H}</math> by
 +
:<math>\tau\mapsto\frac{a\tau+b}{c\tau+d}</math>
 +
for <math>\left ( \begin{array}{cc}a & b \\ c & d \end{array} \right )\in \operatorname{SL}(2,\mathbb{Z})</math>
  
==basic mathematics==
+
;def
* 두 $n\times n$행렬 $A=(a_{ij})$와 $B=(b_{ij})$에 대하여, $AB$의 대각합은 다음과 같이 주어진다
+
A holomorphic function <math>f:\mathbb{H}\to \mathbb{C}</math> is a modular form of weight <math>k</math> (w.r.t. <math>SL(2, \mathbb Z)</math>) if
$$
+
# <math>f \left( \frac{ a\tau +b}{ c\tau + d} \right) = (c\tau +d)^{k} f(\tau)</math>
\operatorname{tr}(AB)=\sum_{i,j=1}^{n}a_{ij}b_{ji}
+
# <math>f</math> is "holomorphic at the cusp", i.e. it has a Fourier expansion of the following form
$$
+
:<math>
* $n=2$인 경우 $\operatorname{tr}(AB)=a_{1,1} b_{1,1}+a_{2,1} b_{1,2}+a_{1,2} b_{2,1}+a_{2,2} b_{2,2}$
+
f(\tau)=\sum_{n=0}^{\infty}a(n)e^{2\pi i n \tau}
* $n=3$인 경우 $\operatorname{tr}(AB)=a_{1,1} b_{1,1}+a_{2,1} b_{1,2}+a_{3,1} b_{1,3}+a_{1,2} b_{2,1}+a_{2,2} b_{2,2}+a_{3,2} b_{2,3}+a_{1,3} b_{3,1}+a_{2,3} b_{3,2}+a_{3,3} b_{3,3}$
+
</math>
  
 +
===Eisenstein series===
 +
* for an integer <math>k\geq 2</math>, define the Eisenstein series by
 +
:<math>
 +
E_{2k}(\tau) : =\frac{1}{2}\sum_{
 +
\substack{
 +
(c,d)\in \mathbb{Z}^2\\
 +
(c,d)=1
 +
}}
 +
\frac{1}{(c\tau+d )^{2k}}
 +
</math>
 +
* Fourier expansion
 +
:<math>E_{2k}(\tau):= 1+\frac {2}{\zeta(1-2k)}\left(\sum_{n=1}^{\infty} \sigma_{2k-1}(n)q^{n} \right)=1-\frac {4k}{B_{2k}}\left(\sum_{n=1}^{\infty} \sigma_{2k-1}(n)q^{n} \right)</math>
 +
where <math>\zeta</math> denotes the Riemann zeta function, <math>B_k</math> Bernoulli number and <math>\sigma_r(n)=\sum_{d|n}d^r</math>
 +
* this is a modular form of weight <math>2k</math>
 +
* for example
 +
:<math>E_4(\tau)= 1+ 240\sum_{n=1}^\infty \sigma_3(n) q^{n}=1 + 240 q + 2160 q^2 + \cdots </math>
 +
:<math>E_6(\tau)=1- 504\sum_{n=1}^\infty \sigma_5(n) q^{n}=1 - 504 q - 16632 q^2 - \cdots </math>
  
==주기 행렬==
+
===the space of modular forms===
* 다음을 만족하는 <math>H_1(X, \mathbb{Z}) \cong \mathbb{Z}^{2g}</math>의 기저, 2g 개의 닫힌 곡선 <math>a_1, \dots, a_g,b_1,\cdots,b_g</math>이 존재 (canonical homology basis)
+
;thm
$$
+
Let <math>M_k</math> be the space of modular forms of weight <math>k</math> and <math>M:=\bigoplus_{k\in \mathbb{Z}_{\geq 0}} M_k</math>. We have
\langle a_i,b_j \rangle = \begin{cases} 1, & \text{if }i=j\\ 0, & \text{if }i\neq j \\ \end{cases}
+
:<math>M=\mathbb{C}[E_4,E_6]</math>
$$
+
* dimension generating function
* 즉, 다음과 같은 intersection form을 가진다
+
:<math>
$$
+
\sum_{k=0}^{\infty}\dim M_k x^k=\frac{1}{\left(1-x^4\right)\left(1-x^{6}\right)}=1+x^4+x^6+x^8+x^{10}+2 x^{12}+x^{14}+2 x^{16}+2 x^{18}+2 x^{20}+\cdots
\begin{array}{c|cc}
+
</math>
\text{} & a&  b  \\
 
\hline
 
a & 0 & I_g \\
 
b & -I_g & 0 
 
\end{array}
 
$$
 
* 다음을 만족하는 <math>H^0(X, K) \cong \mathbb{C}^g</math>의 기저, holomorphic 1-form $\omega_1,\cdots,\omega_{g}$가 존재
 
$$
 
\int_{a_i}\omega_j=\delta_{ij}
 
$$
 
* $\tau_{i,j}=\int_{b_i}\omega_j$로 두면, $\tau=(\tau_{i,j})_{1\leq i,j\leq g}$는 다음의 성질을 만족한다 (리만 겹선형 관계)
 
# $\tau^{\mathrm{T}}=\tau$
 
# $\textrm{Im}(\tau)$는 [[양의 정부호 행렬(positive definite matrix)]]
 
* 즉, $\tau$는 지겔 상반 공간 $\mathcal{H}_g$의 원소이며, $X$의 주기 행렬 (period matrix)라 부른다
 
  
===$g=3$ 인 경우===
+
==theta functions==
$$
+
===notation===
\begin{array}{c|ccc|ccc}
+
* <math>\Lambda\subset \mathbb{R}^n</math> : integral lattice, i.e. a free abelian group with a positive definite symmetric bilinear form, i.e. <math>x\cdot y\in \mathbb{Z}</math> for all <math>x,y\in \Lambda</math>
\text{} & a_1 & a_2 & a_3 & b_1 & b_2 & b_3 \\
+
* we will assume that <math>\Lambda</math> is even, i.e., <math>x\cdot x\in 2\mathbb{Z}</math>
\hline
+
* for a basis of <math>\Lambda</math>, fix <math>M</math>, <math>n\times n</math> matrix whose each row is a basis element
\omega _1 & \left\langle a_1|\omega _1\right\rangle  & \left\langle a_2|\omega _1\right\rangle  & \left\langle a_3|\omega _1\right\rangle  & \left\langle b_1|\omega _1\right\rangle  & \left\langle b_2|\omega _1\right\rangle  & \left\langle b_3|\omega _1\right\rangle  \\
+
* <math>A:=M^tM</math>, Gram matrix of <math>\Lambda</math>
\omega _2 & \left\langle a_1|\omega _2\right\rangle  & \left\langle a_2|\omega _2\right\rangle  & \left\langle a_3|\omega _2\right\rangle  & \left\langle b_1|\omega _2\right\rangle  & \left\langle b_2|\omega _2\right\rangle  & \left\langle b_3|\omega _2\right\rangle  \\
 
\omega _3 & \left\langle a_1|\omega _3\right\rangle  & \left\langle a_2|\omega _3\right\rangle  & \left\langle a_3|\omega _3\right\rangle  & \left\langle b_1|\omega _3\right\rangle  & \left\langle b_2|\omega _3\right\rangle  & \left\langle b_3|\omega _3\right\rangle
 
\end{array}
 
=
 
\begin{array}{c|ccc|ccc}
 
\text{} & a_1 & a_2 & a_3 & b_1 & b_2 & b_3 \\
 
\hline
 
\omega _1 & 1 & 0 & 0 & \tau _{1,1} & \tau _{1,2} & \tau _{1,3} \\
 
\omega _2 & 0 & 1 & 0 & \tau _{2,1} & \tau _{2,2} & \tau _{2,3} \\
 
\omega _3 & 0 & 0 & 1 & \tau _{3,1} & \tau _{3,2} & \tau _{3,3}
 
\end{array}
 
$$
 
여기서 $\left\langle \gamma|\omega\right\rangle=\int_{\gamma}\omega$
 
  
 +
===definition===
 +
* old problem in number theory : find the number of representations of a given integer by the quadratic form associated to <math>\Lambda</math>
 +
* for a given integer <math>N</math>, determine the size of the set <math>\{x\in\Lambda|x\cdot x=2N\}</math> or <math>\{\zeta\in \mathbb{Z}^n|\zeta A \zeta^{t} =2N\}</math>
 +
* denote it by <math>a(N)</math>
 +
* theta function of <math>\Lambda</math> is a holomorphic function on <math>\mathbb{H}</math> given by
 +
:<math>
 +
\Theta_\Lambda(\tau)=\sum_{x\in\Lambda}q^{\frac{x\cdot x}{2}}=\sum_{N=0}^\infty a(N)q^{N},
 +
</math>
 +
where <math>q=e^{2\pi i \tau}</math>
  
==symplectic group==
+
==on theta functions of positive definite even unimodular lattices==
* $M^T J_{n} M = J_{n}$을 만족시키는 $2n\times 2n$ 행렬 $M$ 을 사교행렬이라 함
+
===8차원===
* 여기서 $J_{n}$는 다음과 같이 주어진 $2n\times 2n$ 행렬
+
* <math>\dim M_4=1</math> and thus
$$
+
:<math>\theta_{E_8}(\tau)=E_4(\tau)=1+240 q+2160 q^2+6720 q^3+17520 q^4+30240 q^5+\cdots</math>
J_{n} =\begin{pmatrix}0 & I_n \\-I_n & 0 \\\end{pmatrix}
 
$$
 
* 사교군 $\Gamma_g:={\rm Sp}(2g,\Z)=\{M\in \M_{2g}(\mathbb{Z})|M^T J_{g} M = J_{g}\}$
 
  
 +
===16차원===
 +
* <math>\dim M_8=1</math>, <math>E_8=E_4^2</math> and 
 +
:<math>
 +
\theta_{E_8\oplus E_8}(\tau)=\theta_{D_{16}^{+}}(\tau)=E_8(\tau)\\
 +
E_8(\tau)=1+480 q+61920 q^2+1050240 q^3+7926240 q^4+\cdots
 +
</math>
  
==지겔 상반 공간==
+
===24차원===
* 지겔 상반 공간 $\mathcal{H}_g$
+
* {{수학노트|url=24차원_짝수_자기쌍대_격자}}의 세타함수
$$
+
* modular form of weight 12
\mathcal{H}_g=\left\{\tau \in M_{g \times g}(\mathbb{C}) \ \big| \ \tau^{\mathrm{T}}=\tau, \textrm{Im}(\tau) \text{ positive definite} \right\}
+
* <math>M_{12}=\mathbb{C}\langle E_4^3,E_6^2\rangle</math>
$$
+
* let <math>{\rm gen}(L)</math> be the set of all isomorphim classes of 24-dimensional positive definite even unimodular lattices
* $\mathcal{A}_g=\mathcal{H}_g/\Gamma_g$ : moduli space of principally polarized abelian varieties
+
* to compute <math>\theta_{\Lambda}</math>, find <math>a,b</math> such that <math>\theta_{\Lambda}=a E_4^3+ bE_6^2</math>
 +
* we can easily determine <math>a,b</math> once we know the number <math>r</math> of roots in <math>\Lambda</math> (the coefficient of <math>q</math> in <math>\theta_{\Lambda}</math>) by solving
 +
:<math> \left\{ \begin{array}{c} a+b=1 \\ 720 a - 1008 b=r \end{array} \right. </math>
 +
* weighted average
 +
:<math>\left(\sum_{\Lambda\in {\rm gen}(L)}\frac{\Theta_{\Lambda}(\tau)}{|{\rm Aut}(\Lambda)|}\right)\,\cdot\,
 +
\left(\sum_{\Lambda\in {\rm gen}(L)}\frac{1}{|{\rm Aut}(\Lambda)|}\right)^{-1}=?</math>
 +
* we get
 +
:<math>\left( \sum_{\Lambda\in {\rm gen}(L)}\frac{\Theta_{\Lambda}(\tau)}{|{\rm Aut}(\Lambda)|}\right)\,\cdot\,
 +
\left(\sum_{\Lambda\in {\rm gen}(L)}\frac{1}{|{\rm Aut}(\Lambda)|}\right)^{-1}=E_{12}(\tau)</math>
 +
where <math>E_{12}</math> is the Eisenstein series
 +
:<math>
 +
E_{12}(\tau)=1+\frac{65520 q}{691}+\frac{134250480 q^2}{691}+\frac{11606736960 q^3}{691}+\frac{274945048560 q^4}{691}+\frac{3199218815520 q^5}{691}+\cdots
 +
</math>
  
 +
==Siegel theta series==
 +
* {{수학노트|url=격자의_지겔_세타_급수}}
 +
* for <math>g\in \mathbb{N}</math> and <math>\Lambda</math> of rank <math>n</math>, we will define the Siegel theta series <math>\Theta_\Lambda^{(g)}</math> of degree (or genus) <math>g</math> (<math>g</math> comes from the genus of Riemann surfaces)
 +
* <math>g=1</math> case recovers <math>\Theta_\Lambda^{(1)}=\Theta_\Lambda</math>
 +
;def (half-integral matrix)
 +
A symmetric matrix <math>N\in \operatorname{GL}(g,\mathbb{Q})</math> is called half-integral if <math>2N</math> has integral entries with even integers on the diagonal
 +
===representations of a quadratic form by another quadratic form===
 +
* we want to find the number of representations of a quadratic form by the quadratic form of <math>\Lambda</math>
 +
* let <math>g\leq n</math>
 +
* <math>\underline{x}</math> : <math>g\times n</math> matrix whose row is an element of <math>\Lambda</math>
 +
* for each half-integral <math>g\times g</math> matrix <math>\underline{N}=(N_{ij})</math>, let <math>a(\underline{N})</math> be the number of elements in <math>\{\underline{x}=(x_i)\in\Lambda^{g}| x_i\cdot x_j=2N_{ij}\}</math>
 +
* a given <math>\underline{x}</math> can be written as <math>\underline{x}=\underline{\zeta}M</math> for some <math>\underline{\zeta}</math>, a <math>g\times n</math> integer matrix
 +
* <math>a(\underline{N})</math> is the number of elements in <math>\{\underline{\zeta}\in\mathbb{Z}^{g,n}|\underline{\zeta} A \underline{\zeta}^t =2\underline{N}\}</math>
  
==지겔 모듈라 형식==
+
===definition===
;정의
+
* Let <math>\tau=(\tau_{ij})</math> be a symmetric <math>g\times g</math> matrix
weight이 k이고 genus(또는 degree)가 $g$인 지겔 모듈라 형식은 다음 조건을 만족하는 해석함수 $f:\mathcal{H}_g\to \mathbb{C}$로 정의된다
+
* for <math>\Lambda</math>, the theta series <math>\Theta_\Lambda^{(g)}</math> of genus <math>g</math> is defined by
$$
+
:<math>
f \left( (A\tau +B)(C\tau + D)^{-1}\right) = \det(C\tau +D)^{k} f(\tau),\, \forall \begin{pmatrix}A & B \\ C & D \\\end{pmatrix}\in {\rm Sp}(2g,\Z)
+
\begin{align}
$$
+
\Theta_\Lambda^{(g)}(\tau)&=\sum_{\underline{x}\in\Lambda^{g}}e^{\pi i\operatorname{Tr}(\underline{x}\cdot \underline{x} \tau)}\\
 +
&=\sum_{\underline{\zeta}\in\mathbb{Z}^{g,n}}e^{\pi i\operatorname{Tr}(\underline{\zeta} A \underline{\zeta}^{t}\tau)}\\
 +
&=\sum_{\underline{N}:\text{h.i.}} a(\underline{N})e^{2\pi i\operatorname{Tr}(\underline{N}\tau)}
 +
\end{align} \label{tg}
 +
</math>
  
 +
===note on trace===
 +
* in the last equality, we used the following property of trace
 +
* for two <math>n\times n</math> matrices <math>A=(a_{ij})</math> and <math>B=(b_{ij})</math>,
 +
:<math>
 +
\operatorname{tr}(AB)=\sum_{i,j=1}^{n}a_{ij}b_{ji}
 +
</math>
 +
* if <math>A</math> and <math>B</math> are symmetric,
 +
:<math>
 +
\operatorname{tr}(AB)=\sum_{i,j=1}^{n}a_{ij}b_{ij}
 +
</math>
 +
* the series \ref{tg} converges absolutely if <math>\tau</math> is an element of
 +
:<math>
 +
\mathcal{H}_g:=\left\{\tau \in \operatorname{Mat}_{g \times g}(\mathbb{C}) \ \big| \ \tau^{\mathrm{T}}=\tau, \textrm{Im}(\tau) \text{  positive definite} \right\}
 +
</math>
 +
* it is a holomorphic function on <math>\mathcal{H}_g</math>
  
 +
==Siegel theta functions of even unimodular lattices==
 +
===8차원===
 +
* <math>g=2</math> case
 +
* Fourier coefficient of <math>\Theta_{E_8}^{(2)}</math>
 +
* <math>N = \Bigl( {a \atop b/2} \thinspace {b/2 \atop c} \Bigr) \in
 +
\operatorname{Mat}_{2\times 2}({1 \over 2}\Z)</math>, positive semi-definite, half-integral matrix
 +
* for <math>\tau=\left(
 +
\begin{array}{cc}
 +
\tau _1 & z \\
 +
z & \tau _2
 +
\end{array}
 +
\right)</math>,
 +
:<math>
 +
\operatorname{Tr}(N\tau)=a \tau _1+b z+c \tau _2
 +
</math>
 +
* by setting <math>q_i=e^{2\pi i \tau_i}</math>, <math>\zeta=e^{2\pi i z}</math>, we get
 +
:<math>\exp(2\pi i \operatorname{Tr}(N\tau))=q_1^a\zeta^bq_2^c</math>
 +
* let us compute <math>a(N)</math> for <math>N=
 +
\left(
 +
\begin{array}{cc}
 +
0 & 0 \\
 +
0 & 0
 +
\end{array}
 +
\right), \left(
 +
\begin{array}{cc}
 +
1 & 0 \\
 +
0 & 0
 +
\end{array}
 +
\right), \left(
 +
\begin{array}{cc}
 +
1 & 0 \\
 +
0 & 1
 +
\end{array}
 +
\right)</math>.
 +
* for the third one, we may use the following property of the <math>E_8</math> root system <math>\Phi</math>
 +
# for a given <math>v\in \Phi</math>, there exist 126 elements in <math>\Phi</math> orthogonal to <math>v</math>
 +
# 240*126=30240
 +
* table
 +
:<math>
 +
\begin{array}{c|c|c|c|c|c|c|c|c|c|c}
 +
N & \left(
 +
\begin{array}{cc}
 +
0 & 0 \\
 +
0 & 0
 +
\end{array}
 +
\right) & \left(
 +
\begin{array}{cc}
 +
1 & 0 \\
 +
0 & 0
 +
\end{array}
 +
\right) & \left(
 +
\begin{array}{cc}
 +
0 & 0 \\
 +
0 & 1
 +
\end{array}
 +
\right) & \left(
 +
\begin{array}{cc}
 +
2 & 0 \\
 +
0 & 0
 +
\end{array}
 +
\right) & \left(
 +
\begin{array}{cc}
 +
0 & 0 \\
 +
0 & 2
 +
\end{array}
 +
\right) & \left(
 +
\begin{array}{cc}
 +
1 & -1 \\
 +
-1 & 1
 +
\end{array}
 +
\right) & \left(
 +
\begin{array}{cc}
 +
1 & -\frac{1}{2} \\
 +
-\frac{1}{2} & 1
 +
\end{array}
 +
\right) & \left(
 +
\begin{array}{cc}
 +
1 & 0 \\
 +
0 & 1
 +
\end{array}
 +
\right) & \left(
 +
\begin{array}{cc}
 +
1 & \frac{1}{2} \\
 +
\frac{1}{2} & 1
 +
\end{array}
 +
\right) & \left(
 +
\begin{array}{cc}
 +
1 & 1 \\
 +
1 & 1
 +
\end{array}
 +
\right) \\
 +
\hline
 +
a(N) & 1 & 240 & 240 & 2160 & 2160 & 240 & 13440 & 30240 & 13440 & 240 \\
 +
\hline
 +
\exp(2\pi i \operatorname{Tr}(N\tau)) & 1 & q_1 & q_2 & q_1^2 & q_2^2 & \frac{q_1 q_2}{\zeta^2} & \frac{q_1 q_2}{\zeta} & q_1 q_2 & q_1 q_2 \zeta & q_1 q_2 \zeta^2
 +
\end{array}
 +
</math>
  
==푸리에 전개==
+
===16차원===
* 지겔 모듈라 형식 $f\in M_k(\Gamma_g)$는 다음과 같은 형태의 푸리에 전개를 가진다
+
* <math>E_8\oplus E_8</math> and <math>D_{16}^{+}</math> lattice
$$f(\tau)=\sum_{T}a(T)\exp\left(2\pi i \operatorname{Tr}(T\tau)\right)$$
+
* for <math>g=1,2,3</math>, <math>\Theta_{E_8\oplus E_8}^{(g)}=\Theta_{D_{16}^{+}}^{(g)}</math>
여기서 $T\in \operatorname{Mat}_2(\frac{1}{2}\mathbb{Z})$는 대각성분이 정수인 대칭행렬.
+
* <math>\Theta^{(4)}_{E_8\oplus E_8}\neq \Theta^{(4)}_{D_{16}^{+}}</math>
;Kocher 원리
+
* <math>\Theta^{(4)}_{E_8\oplus E_8}-\Theta^{(4)}_{D_{16}^{+}}</math>, Siegel cusp form of weight 8 called the Schottky form
지겔 모듈라 형식 $f\in M_k(\Gamma_g)$의 푸리에 전개에서, $T$가 positive semi-definite 행렬이 아니면, $a(T)=0$이다
 
  
 +
===24차원===
 +
* for 24 Niemeier lattices, the associated theta series are linearly dependent in degree <math>\leq</math> 11 and linearly independent in degree 12 (Borcherds-Freitag-Weissauer, 1998)
  
==지겔 모듈라 형식의 예==
+
;thm
* [[지겔-아이젠슈타인 급수]]
+
For a positive definite even unimodular lattice <math>\Lambda</math>, <math>\theta^{(g)}_{\Lambda}</math> is a Siegel modular form of weight <math>\frac{n}{2}</math> w.r.t. <math>\Gamma_g</math>
* [[격자의 지겔 세타 급수]]
 
  
 +
==symplectic group==
 +
* symplectic group <math>\Gamma_g:=\operatorname{Sp}(2g,\Z)=\{M\in \operatorname{GL}(2g,\mathbb{Z})|M^T J_{g} M = J_{g}\}</math>
 +
where
 +
:<math>
 +
J_{g} =\begin{pmatrix}0 & I_g \\-I_g & 0 \\\end{pmatrix}
 +
</math>
 +
* <math>2g\times 2g</math> matrix
 +
* one can check that for
 +
:<math>M=\begin{pmatrix}A & B \\ C & D \\\end{pmatrix}\in \Gamma_g,</math>
 +
:<math>
 +
\begin{align}
 +
A^tC=C^tA \\
 +
B^tD=D^tB \\
 +
A^tD-C^tB= I_g
 +
\end{align}
 +
</math>
 +
* the lattice <math>\mathbb{Z}^{2g}</math> of rank <math>2g</math> with basis <math>a_1,\cdots, a_g,b_1\cdots,b_g</math> with the symplectic form
 +
:<math>
 +
\langle a_i,b_j \rangle = \begin{cases} 1, & \text{if }i=j\\ 0, & \text{if }i\neq j \\ \end{cases}
 +
</math>
 +
* then <math>\Gamma_g=\operatorname{Aut}(\mathbb{Z}^{2g},\langle,\rangle)</math>
 +
* note that
 +
:<math>
 +
\begin{pmatrix} I_g & S \\ 0& I_g  \\\end{pmatrix} \in \Gamma_g
 +
</math>
 +
for any symmetric integral matrix <math>S</math>
  
 +
==Siegel upper-half space==
 +
* <math>\mathcal{H}_g</math>
 +
:<math>
 +
\mathcal{H}_g=\left\{\tau \in \operatorname{Mat}_{g \times g}(\mathbb{C}) \ \big| \ \tau^{\mathrm{T}}=\tau, \textrm{Im}(\tau) \text{  positive definite} \right\}
 +
</math>
 +
* there is an action of <math>\Gamma_g</math> on <math>\mathcal{H}_g</math> by
 +
:<math>
 +
\tau\mapsto \gamma(\tau)=(A\tau +B)(C\tau + D)^{-1}
 +
</math>
 +
* we need to check that <math>C\tau + D</math> Is invertible and <math>\Im{\gamma(\tau)}>0 </math>
  
==개요==
+
===Riemann bilinear relation===
* 자연수 $g$와 격자 $\Lambda$에 대하여 정의되는 함수 $\Theta_\Lambda^{(g)}(\tau)$
+
* {{수학노트|url=리만_곡면의_주기_행렬과_겹선형_관계_(bilinear_relation)}}
* 정의역은 지겔 상반 공간
+
* <math>X</math> : compact Riemann surface of genus <math>g</math>
$$
+
* there exists a basis <math>a_1, \dots, a_g,b_1,\cdots,b_g</math> of <math>H_1(X, \mathbb{Z}) \cong \mathbb{Z}^{2g}</math> with the intersection pairing (canonical homology basis)
\mathcal{H}_g=\left\{\tau \in M_{g \times g}(\mathbb{C}) \ \big| \ \tau^{\mathrm{T}}=\tau, \textrm{Im}(\tau) \text{  positive definite} \right\}
+
:<math>
$$
+
\langle a_i,b_j \rangle = \begin{cases} 1, & \text{if }i=j\\ 0, & \text{if }i\neq j \\ \end{cases}
* [[격자의 세타함수]]는 $g=1$인 경우에 해당
+
</math>
===기호===
+
* there exists a basis of the space of holomorphic 1-form, <math>\omega_1,\cdots,\omega_{g}</math> such that
* $\Lambda\subset \mathbb{R}^n$ $n$차원 격자
+
:<math>
* $M$는 각 행이 $\Lambda$의 기저가 되는 $n\times n$ 행렬
+
\int_{a_i}\omega_j=\delta_{ij}
* $A:=M^tM$는 $\Lambda$의 그램 행렬
+
</math>
 +
* if we set <math>\tau_{i,j}=\int_{b_i}\omega_j</math>, then <math>\tau=(\tau_{i,j})_{1\leq i,j\leq g}</math> satisfies the following properties
 +
# <math>\tau^{\mathrm{T}}=\tau</math>
 +
# <math>\textrm{Im}(\tau)</math> is positive definite
 +
* this is called the Riemann bilinear relation
 +
* <math>\tau\in \mathcal{H}_g</math> and and it is called a period matrix of <math>X</math>
 +
* <math>\mathcal{A}_g=\mathcal{H}_g/\Gamma_g</math> : moduli space of principally polarized abelian varieties
  
 +
==Siegel modular forms==
 +
* {{수학노트|url=지겔_모듈라_형식}}
 +
;definition
 +
A holomorphic function <math>f:\mathcal{H}_g\to \mathbb{C}</math> is a Siegel modular form of weight k and genus(or degree) <math>g</math> if
 +
:<math>
 +
f \left( (A\tau +B)(C\tau + D)^{-1}\right) = \det(C\tau +D)^{k} f(\tau),\, \forall \begin{pmatrix}A & B \\ C & D \\\end{pmatrix}\in \Gamma_g
 +
</math>
 +
and it must be holomorphic at the cusp if <math>g=1</math>
 +
* denote the vector space of such functions as <math>M_k(\Gamma_g)</math>
  
==g가 1인 경우==
+
===Fourier expansion===
* 격자 $\Lambda$에 대하여, $N_m$를 $\{x\in\Lambda|x\cdot x=m\}$의 원소의 개수로 정의
+
* note that
* $N_m$는 $\zeta A \zeta^{t} =m$를 만족하는 정수벡터 $\zeta$의 개수로 이해할 수 있다
+
:<math>
* 다시 말해, $\Lambda$에 의해 얻어지는 이차형식이 정수 $m$을 표현하는 방법의 수이다
+
\begin{pmatrix} I_g & S \\ 0& I_g  \\\end{pmatrix}\cdot \tau = \tau+S
* $\Lambda$의 세타함수는 복소상반평면 $\mathcal{H}_1$을 정의역으로 하는 다음과 같은 함수가 된다
+
</math>
$$
+
* <math>f\in M_k(\Gamma_g)</math> satisfies <math>f(\tau+S)=f(\tau)</math> for any symmetric integral <math>S</math>
\Theta_\Lambda(\tau)=\sum_{x\in\Lambda}q^{x\cdot x}=\sum_{m=0}^\infty N_mq^m,
+
* we get the following expansion
$$
+
:<math>
여기서 $q=e^{\pi i \tau}$이고, $\tau\in\mathcal{H}_1$.
+
f(q_{11},\cdots, q_{gg})=\sum_{n_{11},\cdots, n_{ij},\cdots, n_{gg}\in \mathbb{Z}}a(n_{11},\cdots, n_{gg})q_{11}^{n_{11}}\cdots q_{gg}^{n_{gg}} \label{fou1}
===예===
+
</math>
* $\mathbb{Z}$의 세타함수는 자코비 세타함수로 다음과 같다
+
where <math>q_{ij}=e^{2\pi i \tau_{ij}}</math>, <math>i\leq j</math>
$$\Theta_{\mathbb{Z}}(\tau)=\sum_{m=-\infty}^{\infty}q^{m^2}=1+2q+2q^4+2q^9+\cdots\equiv\theta[{}_0^0](\tau)$$
+
* define a symmetric matrix <math>N=(N_{ij})_{1\leq i,j\leq g}</math> as
 +
:<math>
 +
N_{ij}=
 +
\begin{cases}
 +
n_{ii}, & \text{if </math>i=j<math>}\\
 +
n_{ij}/2, & \text{if </math>i\neq j<math>}
 +
\end{cases}
 +
</math>
 +
* <math>\operatorname{Tr}(N\tau)=\sum_{i=1}^{g}N_{ii}\tau_{ii}+2\sum_{1\leq i<j\leq g}N_{ij}\tau_{ij}</math>
 +
* <math>\exp(2\pi i \operatorname{Tr}(N\tau))=q_{11}^{n_{11}}\cdots q_{gg}^{n_{gg}}</math>
 +
* \ref{fou1} can be rewritten as
 +
:<math>f(\tau)=\sum_{N}a(N)\exp\left(2\pi i \operatorname{Tr}(N\tau)\right)</math>
 +
where the summation is over <math>N=(N_{ij})\in \operatorname{Mat}_g(\frac{1}{2}\mathbb{Z})</math> half-integral matrix
 +
;Koecher Principle
 +
For a Siegel modular form <math>f\in M_k(\Gamma_g)</math>, if <math>N</math> is not a positive semi-definite matrix, then <math>a(N)=0</math>. (this is why holomorphicity at the cusp is not necessary if <math>g>1</math>)
  
 +
==지겔 모듈라 형식의 예==
 +
* [[격자의 지겔 세타 급수]]
 +
* {{수학노트|url=지겔-아이젠슈타인_급수}}
 +
:<math>
 +
E_{k}^{(g)}(\tau) = \sum_{(C,D)} \frac{1}{\det(C\tau +D)^{k}}
 +
</math>
 +
where the summation is over all
 +
:<math>
 +
\begin{pmatrix}A & B \\ C & D \\\end{pmatrix}\in \Gamma_{g,0}\backslash \Gamma_{g}
 +
</math>
 +
and
 +
:<math>
 +
\Gamma_{g,0}=\{\begin{pmatrix}A & B \\ 0 & D \\\end{pmatrix}\in \Gamma_{g}\}
 +
</math>
 +
(the summation extends over all classes of coprime symmetric pairs, i. e. over all inequivalent bottom rows of elements of <math>\Gamma_g</math> with respect to left multiplications by unimodular integer matrices of degree <math>g</math>. In other words, the sum is over a full set of representatives for the cosets <math>\operatorname{GL}(g,\mathbb{Z})\backslash \Gamma_{g}</math>)
 +
* [[Fourier coefficients of Siegel-Eisenstein series]]
  
==일반적인 자연수 g에 대한 세타함수의 일반화==
+
==Siegel-Weil formula==
* 자연수 $g$, (리만곡면의 genus에서 g가 온 것이다)
+
* [[Siegel-Weil formula]]
* $\underline{\zeta}$를 정수계수를 갖는 $g\times n$ 행렬
+
* {{수학노트|url=지겔-베유_공식}}
* $\underline{x}$는 각 행이 격자 $\Lambda$의 원소가 되는 $g\times n$ 행렬이라 하자
+
;thm
* 주어진 $\underline{x}$는 적당한 $\underline{\zeta}$에 대하여 $\underline{x}=\underline{\zeta}M$꼴로 쓰여진다
+
For a positive definite even unimodular lattice <math>L</math>,
* 이제 각각의 정수계수 $g\times g$ 행렬 $\underline{m}$에 대하여, $N_{\underline{m}}\in\mathbb{Z}$를 다음 방정식의 해의 개수로 정의하자
+
:<math>\left( \sum_{M\in {\rm gen}(L)}\frac{\Theta_M^{(g)}(Z)}{|{\rm Aut}(M)|}\right)\,\cdot\,
$$
+
\left(\sum_{M\in {\rm gen}(L)}\frac{1}{|{\rm Aut}(M)|}\right)^{-1}=
\underline{\zeta} A \underline{\zeta}^t =\underline{m},
+
E^{(g)}_{k}(Z),</math>
$$
 
* $\underline{m}$의 성분 $m_{i,j}$는 $\underline{x}$의 두 행 $x_i,x_j\in\Lambda$ 사이의 내적이다
 
* 따라서 $N_{\underline{m}}$는 $x_i\cdot x_j=m_{ij}$를 만족하는 $\underline{x}=(x_i)$의 개수이다
 
** $N_{\underline{m}}$은 $\underline{m}$에 대응되는 이차형식을 $\Lambda$에 대응되는 이차형식으로 표현하는 방법의 개수로 이해할 수 있다
 
* $\Lambda$에 대한 genus $g$ 세타함수는 지겔 상반 공간 $\mathcal{H}_g$에 정의되는 다음과 같은 해석함수이다
 
$$
 
\begin{align}
 
\Theta_\Lambda^{(g)}(\tau)&=\sum_{x\in\Lambda^{(g)}}e^{\pi i\operatorname{Tr}(x\cdot x)}\\
 
&=\sum_{\underline{\zeta}\in\mathbb{Z}^{g,n}}e^{\pi i\operatorname{Tr}(\underline{\zeta} A \underline{\zeta}^{t}\tau)}\\
 
&=\sum_{\underline{m}} N_{\underline{m}}\prod_{i\leq j}e^{\pi i m_{ij}\tau_{ij}},
 
\end{align}
 
$$
 
* 마지막 등식에서는 [[행렬의 대각합 (trace)]]의 성질이 사용되었다
 
* 짝수 자기쌍대 격자의 세타함수는 $\Gamma_g$에 대하여 weight $\frac{n}{2}$인 지겔 모듈라 형식
 
* 홀수 자기쌍대 격자는 $\Gamma_g(1,2)$에 대하여 weight $\frac{n}{2}$인 지겔 모듈라 형식
 
* 여기서
 
\begin{align*}
 
& \Gamma_g:={\rm Sp}(2g,\mathbb{Z}), \qquad \Gamma_g(2):=\{ M\in \Gamma_g \mid M=\mathbb{I}\ {\rm mod} 2 \} \\
 
& \Gamma_g (1,2) :=\{ M=\left(^A_C{}^B_D \right) \in \Gamma_g (2) \mid \ A B^{t} ={\rm diag} C\ D^{t} =0 \ {\rm mod} 2 \}.
 
\end{align*}
 
  
 +
Moreover, the Fourier coefficients <math>a_{E}(N)</math> of <math>E</math> can be expressed as an infinite product of [[Local density of quadratic form|local densities]]
 +
:<math>
 +
a_{E}(N)=\prod_{p:\text{primes}}\beta_{L,p}(N) \label{lp}
 +
</math>
 +
===mass formula===
 +
* for a half-integral <math>N</math>,
 +
:<math>
 +
a_{E}(N)=\left( \sum_{M\in {\rm gen}(L)}\frac{r_M(N)}{|{\rm Aut}(M)|}\right)\,\cdot\,
 +
\left(\sum_{M\in {\rm gen}(L)}\frac{1}{|{\rm Aut}(M)|}\right)^{-1}
 +
</math>
 +
where <math>\Theta_M^{(g)}(Z)=\sum_{N}r_M(N)\exp\left(2\pi i \operatorname{Tr}(N\tau)\right)</math>
 +
* if <math>2N</math> is a Gram matrix of <math>L</math>, then we obtain
 +
:<math>
 +
a_{E}(N)=\left(\sum_{M\in {\rm gen}(L)}\frac{1}{|{\rm Aut}(M)|}\right)^{-1}
 +
</math>
 +
as
 +
:<math>
 +
r_M(N) = \begin{cases} |\operatorname{Aut}(L)|, & \text{if }L\sim M \\ 0, & \text{if }L\nsim M \\ \end{cases}
 +
</math>
 +
* then we can express
 +
:<math>
 +
a_{E}(N)=\left(\sum_{M\in {\rm gen}(L)}\frac{1}{|{\rm Aut}(M)|}\right)^{-1}
 +
</math>
 +
in terms of local densities \ref{lp}, which gives the Smith-Minkowski-Siegel mass formula
  
==related items==
 
* [[Siegel-Weil formula]]
 
  
 
[[분류:talks and lecture notes]]
 
[[분류:talks and lecture notes]]
 +
[[분류:theta]]
 +
[[분류:migrate]]

2020년 11월 13일 (금) 03:18 기준 최신판

overview

  • Siegel theta series
  • Siegel modular forms
  • Siegel-Weil formula

modular forms

  • \(\mathbb{H}=\{\tau\in \mathbb{C}|\Im \tau>0\}\)
  • modular group \(\Gamma=SL(2, \mathbb Z) = \left \{ \left. \left ( \begin{array}{cc}a & b \\ c & d \end{array} \right )\right| a, b, c, d \in \mathbb Z,\ ad-bc = 1 \right \}\)
  • \(\operatorname{PSL}(2,\mathbb{Z})=\operatorname{SL}(2,\mathbb{Z})/\{\pm I\}\) acts on \(\mathbb{H}\) by

\[\tau\mapsto\frac{a\tau+b}{c\tau+d}\] for \(\left ( \begin{array}{cc}a & b \\ c & d \end{array} \right )\in \operatorname{SL}(2,\mathbb{Z})\)

def

A holomorphic function \(f:\mathbb{H}\to \mathbb{C}\) is a modular form of weight \(k\) (w.r.t. \(SL(2, \mathbb Z)\)) if

  1. \(f \left( \frac{ a\tau +b}{ c\tau + d} \right) = (c\tau +d)^{k} f(\tau)\)
  2. \(f\) is "holomorphic at the cusp", i.e. it has a Fourier expansion of the following form

\[ f(\tau)=\sum_{n=0}^{\infty}a(n)e^{2\pi i n \tau} \]

Eisenstein series

  • for an integer \(k\geq 2\), define the Eisenstein series by

\[ E_{2k}(\tau) : =\frac{1}{2}\sum_{ \substack{ (c,d)\in \mathbb{Z}^2\\ (c,d)=1 }} \frac{1}{(c\tau+d )^{2k}} \]

  • Fourier expansion

\[E_{2k}(\tau):= 1+\frac {2}{\zeta(1-2k)}\left(\sum_{n=1}^{\infty} \sigma_{2k-1}(n)q^{n} \right)=1-\frac {4k}{B_{2k}}\left(\sum_{n=1}^{\infty} \sigma_{2k-1}(n)q^{n} \right)\] where \(\zeta\) denotes the Riemann zeta function, \(B_k\) Bernoulli number and \(\sigma_r(n)=\sum_{d|n}d^r\)

  • this is a modular form of weight \(2k\)
  • for example

\[E_4(\tau)= 1+ 240\sum_{n=1}^\infty \sigma_3(n) q^{n}=1 + 240 q + 2160 q^2 + \cdots \] \[E_6(\tau)=1- 504\sum_{n=1}^\infty \sigma_5(n) q^{n}=1 - 504 q - 16632 q^2 - \cdots \]

the space of modular forms

thm

Let \(M_k\) be the space of modular forms of weight \(k\) and \(M:=\bigoplus_{k\in \mathbb{Z}_{\geq 0}} M_k\). We have \[M=\mathbb{C}[E_4,E_6]\]

  • dimension generating function

\[ \sum_{k=0}^{\infty}\dim M_k x^k=\frac{1}{\left(1-x^4\right)\left(1-x^{6}\right)}=1+x^4+x^6+x^8+x^{10}+2 x^{12}+x^{14}+2 x^{16}+2 x^{18}+2 x^{20}+\cdots \]

theta functions

notation

  • \(\Lambda\subset \mathbb{R}^n\) : integral lattice, i.e. a free abelian group with a positive definite symmetric bilinear form, i.e. \(x\cdot y\in \mathbb{Z}\) for all \(x,y\in \Lambda\)
  • we will assume that \(\Lambda\) is even, i.e., \(x\cdot x\in 2\mathbb{Z}\)
  • for a basis of \(\Lambda\), fix \(M\), \(n\times n\) matrix whose each row is a basis element
  • \(A:=M^tM\), Gram matrix of \(\Lambda\)

definition

  • old problem in number theory : find the number of representations of a given integer by the quadratic form associated to \(\Lambda\)
  • for a given integer \(N\), determine the size of the set \(\{x\in\Lambda|x\cdot x=2N\}\) or \(\{\zeta\in \mathbb{Z}^n|\zeta A \zeta^{t} =2N\}\)
  • denote it by \(a(N)\)
  • theta function of \(\Lambda\) is a holomorphic function on \(\mathbb{H}\) given by

\[ \Theta_\Lambda(\tau)=\sum_{x\in\Lambda}q^{\frac{x\cdot x}{2}}=\sum_{N=0}^\infty a(N)q^{N}, \] where \(q=e^{2\pi i \tau}\)

on theta functions of positive definite even unimodular lattices

8차원

  • \(\dim M_4=1\) and thus

\[\theta_{E_8}(\tau)=E_4(\tau)=1+240 q+2160 q^2+6720 q^3+17520 q^4+30240 q^5+\cdots\]

16차원

  • \(\dim M_8=1\), \(E_8=E_4^2\) and

\[ \theta_{E_8\oplus E_8}(\tau)=\theta_{D_{16}^{+}}(\tau)=E_8(\tau)\\ E_8(\tau)=1+480 q+61920 q^2+1050240 q^3+7926240 q^4+\cdots \]

24차원

  • 틀:수학노트의 세타함수
  • modular form of weight 12
  • \(M_{12}=\mathbb{C}\langle E_4^3,E_6^2\rangle\)
  • let \({\rm gen}(L)\) be the set of all isomorphim classes of 24-dimensional positive definite even unimodular lattices
  • to compute \(\theta_{\Lambda}\), find \(a,b\) such that \(\theta_{\Lambda}=a E_4^3+ bE_6^2\)
  • we can easily determine \(a,b\) once we know the number \(r\) of roots in \(\Lambda\) (the coefficient of \(q\) in \(\theta_{\Lambda}\)) by solving

\[ \left\{ \begin{array}{c} a+b=1 \\ 720 a - 1008 b=r \end{array} \right. \]

  • weighted average

\[\left(\sum_{\Lambda\in {\rm gen}(L)}\frac{\Theta_{\Lambda}(\tau)}{|{\rm Aut}(\Lambda)|}\right)\,\cdot\, \left(\sum_{\Lambda\in {\rm gen}(L)}\frac{1}{|{\rm Aut}(\Lambda)|}\right)^{-1}=?\]

  • we get

\[\left( \sum_{\Lambda\in {\rm gen}(L)}\frac{\Theta_{\Lambda}(\tau)}{|{\rm Aut}(\Lambda)|}\right)\,\cdot\, \left(\sum_{\Lambda\in {\rm gen}(L)}\frac{1}{|{\rm Aut}(\Lambda)|}\right)^{-1}=E_{12}(\tau)\] where \(E_{12}\) is the Eisenstein series \[ E_{12}(\tau)=1+\frac{65520 q}{691}+\frac{134250480 q^2}{691}+\frac{11606736960 q^3}{691}+\frac{274945048560 q^4}{691}+\frac{3199218815520 q^5}{691}+\cdots \]

Siegel theta series

  • 틀:수학노트
  • for \(g\in \mathbb{N}\) and \(\Lambda\) of rank \(n\), we will define the Siegel theta series \(\Theta_\Lambda^{(g)}\) of degree (or genus) \(g\) (\(g\) comes from the genus of Riemann surfaces)
  • \(g=1\) case recovers \(\Theta_\Lambda^{(1)}=\Theta_\Lambda\)
def (half-integral matrix)

A symmetric matrix \(N\in \operatorname{GL}(g,\mathbb{Q})\) is called half-integral if \(2N\) has integral entries with even integers on the diagonal

representations of a quadratic form by another quadratic form

  • we want to find the number of representations of a quadratic form by the quadratic form of \(\Lambda\)
  • let \(g\leq n\)
  • \(\underline{x}\) \[g\times n\] matrix whose row is an element of \(\Lambda\)
  • for each half-integral \(g\times g\) matrix \(\underline{N}=(N_{ij})\), let \(a(\underline{N})\) be the number of elements in \(\{\underline{x}=(x_i)\in\Lambda^{g}| x_i\cdot x_j=2N_{ij}\}\)
  • a given \(\underline{x}\) can be written as \(\underline{x}=\underline{\zeta}M\) for some \(\underline{\zeta}\), a \(g\times n\) integer matrix
  • \(a(\underline{N})\) is the number of elements in \(\{\underline{\zeta}\in\mathbb{Z}^{g,n}|\underline{\zeta} A \underline{\zeta}^t =2\underline{N}\}\)

definition

  • Let \(\tau=(\tau_{ij})\) be a symmetric \(g\times g\) matrix
  • for \(\Lambda\), the theta series \(\Theta_\Lambda^{(g)}\) of genus \(g\) is defined by

\[ \begin{align} \Theta_\Lambda^{(g)}(\tau)&=\sum_{\underline{x}\in\Lambda^{g}}e^{\pi i\operatorname{Tr}(\underline{x}\cdot \underline{x} \tau)}\\ &=\sum_{\underline{\zeta}\in\mathbb{Z}^{g,n}}e^{\pi i\operatorname{Tr}(\underline{\zeta} A \underline{\zeta}^{t}\tau)}\\ &=\sum_{\underline{N}:\text{h.i.}} a(\underline{N})e^{2\pi i\operatorname{Tr}(\underline{N}\tau)} \end{align} \label{tg} \]

note on trace

  • in the last equality, we used the following property of trace
  • for two \(n\times n\) matrices \(A=(a_{ij})\) and \(B=(b_{ij})\),

\[ \operatorname{tr}(AB)=\sum_{i,j=1}^{n}a_{ij}b_{ji} \]

  • if \(A\) and \(B\) are symmetric,

\[ \operatorname{tr}(AB)=\sum_{i,j=1}^{n}a_{ij}b_{ij} \]

  • the series \ref{tg} converges absolutely if \(\tau\) is an element of

\[ \mathcal{H}_g:=\left\{\tau \in \operatorname{Mat}_{g \times g}(\mathbb{C}) \ \big| \ \tau^{\mathrm{T}}=\tau, \textrm{Im}(\tau) \text{ positive definite} \right\} \]

  • it is a holomorphic function on \(\mathcal{H}_g\)

Siegel theta functions of even unimodular lattices

8차원

  • \(g=2\) case
  • Fourier coefficient of \(\Theta_{E_8}^{(2)}\)
  • \(N = \Bigl( {a \atop b/2} \thinspace {b/2 \atop c} \Bigr) \in \operatorname{Mat}_{2\times 2}({1 \over 2}\Z)\), positive semi-definite, half-integral matrix
  • for \(\tau=\left( \begin{array}{cc} \tau _1 & z \\ z & \tau _2 \end{array} \right)\),

\[ \operatorname{Tr}(N\tau)=a \tau _1+b z+c \tau _2 \]

  • by setting \(q_i=e^{2\pi i \tau_i}\), \(\zeta=e^{2\pi i z}\), we get

\[\exp(2\pi i \operatorname{Tr}(N\tau))=q_1^a\zeta^bq_2^c\]

  • let us compute \(a(N)\) for \(N= \left( \begin{array}{cc} 0 & 0 \\ 0 & 0 \end{array} \right), \left( \begin{array}{cc} 1 & 0 \\ 0 & 0 \end{array} \right), \left( \begin{array}{cc} 1 & 0 \\ 0 & 1 \end{array} \right)\).
  • for the third one, we may use the following property of the \(E_8\) root system \(\Phi\)
  1. for a given \(v\in \Phi\), there exist 126 elements in \(\Phi\) orthogonal to \(v\)
  2. 240*126=30240
  • table

\[ \begin{array}{c|c|c|c|c|c|c|c|c|c|c} N & \left( \begin{array}{cc} 0 & 0 \\ 0 & 0 \end{array} \right) & \left( \begin{array}{cc} 1 & 0 \\ 0 & 0 \end{array} \right) & \left( \begin{array}{cc} 0 & 0 \\ 0 & 1 \end{array} \right) & \left( \begin{array}{cc} 2 & 0 \\ 0 & 0 \end{array} \right) & \left( \begin{array}{cc} 0 & 0 \\ 0 & 2 \end{array} \right) & \left( \begin{array}{cc} 1 & -1 \\ -1 & 1 \end{array} \right) & \left( \begin{array}{cc} 1 & -\frac{1}{2} \\ -\frac{1}{2} & 1 \end{array} \right) & \left( \begin{array}{cc} 1 & 0 \\ 0 & 1 \end{array} \right) & \left( \begin{array}{cc} 1 & \frac{1}{2} \\ \frac{1}{2} & 1 \end{array} \right) & \left( \begin{array}{cc} 1 & 1 \\ 1 & 1 \end{array} \right) \\ \hline a(N) & 1 & 240 & 240 & 2160 & 2160 & 240 & 13440 & 30240 & 13440 & 240 \\ \hline \exp(2\pi i \operatorname{Tr}(N\tau)) & 1 & q_1 & q_2 & q_1^2 & q_2^2 & \frac{q_1 q_2}{\zeta^2} & \frac{q_1 q_2}{\zeta} & q_1 q_2 & q_1 q_2 \zeta & q_1 q_2 \zeta^2 \end{array} \]

16차원

  • \(E_8\oplus E_8\) and \(D_{16}^{+}\) lattice
  • for \(g=1,2,3\), \(\Theta_{E_8\oplus E_8}^{(g)}=\Theta_{D_{16}^{+}}^{(g)}\)
  • \(\Theta^{(4)}_{E_8\oplus E_8}\neq \Theta^{(4)}_{D_{16}^{+}}\)
  • \(\Theta^{(4)}_{E_8\oplus E_8}-\Theta^{(4)}_{D_{16}^{+}}\), Siegel cusp form of weight 8 called the Schottky form

24차원

  • for 24 Niemeier lattices, the associated theta series are linearly dependent in degree \(\leq\) 11 and linearly independent in degree 12 (Borcherds-Freitag-Weissauer, 1998)
thm

For a positive definite even unimodular lattice \(\Lambda\), \(\theta^{(g)}_{\Lambda}\) is a Siegel modular form of weight \(\frac{n}{2}\) w.r.t. \(\Gamma_g\)

symplectic group

  • symplectic group \(\Gamma_g:=\operatorname{Sp}(2g,\Z)=\{M\in \operatorname{GL}(2g,\mathbb{Z})|M^T J_{g} M = J_{g}\}\)

where \[ J_{g} =\begin{pmatrix}0 & I_g \\-I_g & 0 \\\end{pmatrix} \]

  • \(2g\times 2g\) matrix
  • one can check that for

\[M=\begin{pmatrix}A & B \\ C & D \\\end{pmatrix}\in \Gamma_g,\] \[ \begin{align} A^tC=C^tA \\ B^tD=D^tB \\ A^tD-C^tB= I_g \end{align} \]

  • the lattice \(\mathbb{Z}^{2g}\) of rank \(2g\) with basis \(a_1,\cdots, a_g,b_1\cdots,b_g\) with the symplectic form

\[ \langle a_i,b_j \rangle = \begin{cases} 1, & \text{if }i=j\\ 0, & \text{if }i\neq j \\ \end{cases} \]

  • then \(\Gamma_g=\operatorname{Aut}(\mathbb{Z}^{2g},\langle,\rangle)\)
  • note that

\[ \begin{pmatrix} I_g & S \\ 0& I_g \\\end{pmatrix} \in \Gamma_g \] for any symmetric integral matrix \(S\)

Siegel upper-half space

  • \(\mathcal{H}_g\)

\[ \mathcal{H}_g=\left\{\tau \in \operatorname{Mat}_{g \times g}(\mathbb{C}) \ \big| \ \tau^{\mathrm{T}}=\tau, \textrm{Im}(\tau) \text{ positive definite} \right\} \]

  • there is an action of \(\Gamma_g\) on \(\mathcal{H}_g\) by

\[ \tau\mapsto \gamma(\tau)=(A\tau +B)(C\tau + D)^{-1} \]

  • we need to check that \(C\tau + D\) Is invertible and \(\Im{\gamma(\tau)}>0 \)

Riemann bilinear relation

  • 틀:수학노트
  • \(X\) : compact Riemann surface of genus \(g\)
  • there exists a basis \(a_1, \dots, a_g,b_1,\cdots,b_g\) of \(H_1(X, \mathbb{Z}) \cong \mathbb{Z}^{2g}\) with the intersection pairing (canonical homology basis)

\[ \langle a_i,b_j \rangle = \begin{cases} 1, & \text{if }i=j\\ 0, & \text{if }i\neq j \\ \end{cases} \]

  • there exists a basis of the space of holomorphic 1-form, \(\omega_1,\cdots,\omega_{g}\) such that

\[ \int_{a_i}\omega_j=\delta_{ij} \]

  • if we set \(\tau_{i,j}=\int_{b_i}\omega_j\), then \(\tau=(\tau_{i,j})_{1\leq i,j\leq g}\) satisfies the following properties
  1. \(\tau^{\mathrm{T}}=\tau\)
  2. \(\textrm{Im}(\tau)\) is positive definite
  • this is called the Riemann bilinear relation
  • \(\tau\in \mathcal{H}_g\) and and it is called a period matrix of \(X\)
  • \(\mathcal{A}_g=\mathcal{H}_g/\Gamma_g\) : moduli space of principally polarized abelian varieties

Siegel modular forms

definition

A holomorphic function \(f:\mathcal{H}_g\to \mathbb{C}\) is a Siegel modular form of weight k and genus(or degree) \(g\) if \[ f \left( (A\tau +B)(C\tau + D)^{-1}\right) = \det(C\tau +D)^{k} f(\tau),\, \forall \begin{pmatrix}A & B \\ C & D \\\end{pmatrix}\in \Gamma_g \] and it must be holomorphic at the cusp if \(g=1\)

  • denote the vector space of such functions as \(M_k(\Gamma_g)\)

Fourier expansion

  • note that

\[ \begin{pmatrix} I_g & S \\ 0& I_g \\\end{pmatrix}\cdot \tau = \tau+S \]

  • \(f\in M_k(\Gamma_g)\) satisfies \(f(\tau+S)=f(\tau)\) for any symmetric integral \(S\)
  • we get the following expansion

\[ f(q_{11},\cdots, q_{gg})=\sum_{n_{11},\cdots, n_{ij},\cdots, n_{gg}\in \mathbb{Z}}a(n_{11},\cdots, n_{gg})q_{11}^{n_{11}}\cdots q_{gg}^{n_{gg}} \label{fou1} \] where \(q_{ij}=e^{2\pi i \tau_{ij}}\), \(i\leq j\)

  • define a symmetric matrix \(N=(N_{ij})_{1\leq i,j\leq g}\) as

\[ N_{ij}= \begin{cases} n_{ii}, & \text{if \]i=j\(}\\ n_{ij}/2, & \text{if \)i\neq j\(} \end{cases} \)

  • \(\operatorname{Tr}(N\tau)=\sum_{i=1}^{g}N_{ii}\tau_{ii}+2\sum_{1\leq i<j\leq g}N_{ij}\tau_{ij}\)
  • \(\exp(2\pi i \operatorname{Tr}(N\tau))=q_{11}^{n_{11}}\cdots q_{gg}^{n_{gg}}\)
  • \ref{fou1} can be rewritten as

\[f(\tau)=\sum_{N}a(N)\exp\left(2\pi i \operatorname{Tr}(N\tau)\right)\] where the summation is over \(N=(N_{ij})\in \operatorname{Mat}_g(\frac{1}{2}\mathbb{Z})\) half-integral matrix

Koecher Principle

For a Siegel modular form \(f\in M_k(\Gamma_g)\), if \(N\) is not a positive semi-definite matrix, then \(a(N)=0\). (this is why holomorphicity at the cusp is not necessary if \(g>1\))

지겔 모듈라 형식의 예

\[ E_{k}^{(g)}(\tau) = \sum_{(C,D)} \frac{1}{\det(C\tau +D)^{k}} \] where the summation is over all \[ \begin{pmatrix}A & B \\ C & D \\\end{pmatrix}\in \Gamma_{g,0}\backslash \Gamma_{g} \] and \[ \Gamma_{g,0}=\{\begin{pmatrix}A & B \\ 0 & D \\\end{pmatrix}\in \Gamma_{g}\} \] (the summation extends over all classes of coprime symmetric pairs, i. e. over all inequivalent bottom rows of elements of \(\Gamma_g\) with respect to left multiplications by unimodular integer matrices of degree \(g\). In other words, the sum is over a full set of representatives for the cosets \(\operatorname{GL}(g,\mathbb{Z})\backslash \Gamma_{g}\))

Siegel-Weil formula

thm

For a positive definite even unimodular lattice \(L\), \[\left( \sum_{M\in {\rm gen}(L)}\frac{\Theta_M^{(g)}(Z)}{|{\rm Aut}(M)|}\right)\,\cdot\, \left(\sum_{M\in {\rm gen}(L)}\frac{1}{|{\rm Aut}(M)|}\right)^{-1}= E^{(g)}_{k}(Z),\]

Moreover, the Fourier coefficients \(a_{E}(N)\) of \(E\) can be expressed as an infinite product of local densities \[ a_{E}(N)=\prod_{p:\text{primes}}\beta_{L,p}(N) \label{lp} \]

mass formula

  • for a half-integral \(N\),

\[ a_{E}(N)=\left( \sum_{M\in {\rm gen}(L)}\frac{r_M(N)}{|{\rm Aut}(M)|}\right)\,\cdot\, \left(\sum_{M\in {\rm gen}(L)}\frac{1}{|{\rm Aut}(M)|}\right)^{-1} \] where \(\Theta_M^{(g)}(Z)=\sum_{N}r_M(N)\exp\left(2\pi i \operatorname{Tr}(N\tau)\right)\)

  • if \(2N\) is a Gram matrix of \(L\), then we obtain

\[ a_{E}(N)=\left(\sum_{M\in {\rm gen}(L)}\frac{1}{|{\rm Aut}(M)|}\right)^{-1} \] as \[ r_M(N) = \begin{cases} |\operatorname{Aut}(L)|, & \text{if }L\sim M \\ 0, & \text{if }L\nsim M \\ \end{cases} \]

  • then we can express

\[ a_{E}(N)=\left(\sum_{M\in {\rm gen}(L)}\frac{1}{|{\rm Aut}(M)|}\right)^{-1} \] in terms of local densities \ref{lp}, which gives the Smith-Minkowski-Siegel mass formula