From afd5a3bf4c0f72e73277fadcd2aed7691186f719 Mon Sep 17 00:00:00 2001
From: Trance-0 <60459821+Trance-0@users.noreply.github.com>
Date: Tue, 30 Sep 2025 14:22:24 -0500
Subject: [PATCH] updates
---
content/CSE510/_meta.js | 1 +
content/CSE5313/CSE5313_L10.md | 232 +++++++++++++++++++++++++++++++++
content/CSE5313/_meta.js | 1 +
3 files changed, 234 insertions(+)
create mode 100644 content/CSE5313/CSE5313_L10.md
diff --git a/content/CSE510/_meta.js b/content/CSE510/_meta.js
index 3a5efda..aa4ef7e 100644
--- a/content/CSE510/_meta.js
+++ b/content/CSE510/_meta.js
@@ -13,4 +13,5 @@ export default {
CSE510_L8: "CSE510 Deep Reinforcement Learning (Lecture 8)",
CSE510_L9: "CSE510 Deep Reinforcement Learning (Lecture 9)",
CSE510_L10: "CSE510 Deep Reinforcement Learning (Lecture 10)",
+ CSE510_L11: "CSE510 Deep Reinforcement Learning (Lecture 11)",
}
\ No newline at end of file
diff --git a/content/CSE5313/CSE5313_L10.md b/content/CSE5313/CSE5313_L10.md
new file mode 100644
index 0000000..db7ce6c
--- /dev/null
+++ b/content/CSE5313/CSE5313_L10.md
@@ -0,0 +1,232 @@
+# CSE5313 Coding and information theory for data science (Recitation 10)
+
+## Question 2
+
+Let $C$ be a Reed-Solomon code generated by
+
+$$
+G=\begin{bmatrix}
+1 & 1 & \cdots & 1\\
+\alpha_1 & \alpha_2 & \cdots & \alpha_n\\
+\alpha_1^2 & \alpha_2^2 & \cdots & \alpha_n^2\\
+\vdots & \vdots & \cdots & \vdots\\
+\alpha_1^{k-1} & \alpha_2^{k-1} & \cdots & \alpha_n^{k-1}
+\end{bmatrix}
+$$
+
+prove that there exists $v_1,v_2,\ldots,v_n\in \mathbb{F}_q\setminus \{0\}$ such that the parity check matrix is
+
+$$
+H=\begin{bmatrix}
+1 & 1 & \cdots & 1\\
+\alpha_1 & \alpha_2 & \cdots & \alpha_n\\
+\alpha_1^2 & \alpha_2^2 & \cdots & \alpha_n^2\\
+\vdots & \vdots & \cdots & \vdots\\
+\alpha_1^{k-1} & \alpha_2^{k-1} & \cdots & \alpha_n^{k-1}
+\end{bmatrix}\begin{bmatrix}
+v_1 & 0 & \cdots & 0\\
+0 & v_2 & \cdots & 0\\
+\vdots & \vdots & \ddots & \vdots\\
+0 & 0 & \cdots & v_n
+\end{bmatrix}
+$$
+
+### Some lemmas for linear codes
+
+First we introduce the following lemmas for linear codes
+
+Let $G$ and $H$ be the generator and parity-check matrices of (any) linear code
+
+#### Lemma 1
+
+$$
+H G^T = 0
+$$
+
+
+Proof
+
+By definition of generator matrix and parity-check matrix, $forall e_i\in H$, $e_iG^T=0$.
+
+So $H G^T = 0$.
+
+
+#### Lemma 2
+
+Any matrix $M\in \mathbb{F}_q^{(n-k)\times n}$ such that $\operatorname{rank}(M) = n - k$ and $M G^T = 0$ is a parity-check matrix for $C$ (i.e. $C = \ker M$).
+
+
+Proof
+
+It is sufficient to show that the two statements
+
+1. $\forall c\in C, c=uG, u\in \mathbb{F}^k$
+
+$M c^T = M(uG)^T = M(G^T u^T) = 0$ since $M G^T = 0$.
+
+Thus $C \subseteq \ker M$.
+
+2. $\dim (\ker M) +\operatorname{rank}(M) = n$
+
+We proceed by showing that $\dim (\ker M) =\dim (C)$.
+
+Suppose $C$ does not span $\ker M$. Let $u_1,...,u_k$ be a basis for $C$. Then there exists $v\in \ker M\setminus C$.
+
+By linear independence, if we have scalar $a_1,...,a_k$ and $b$ such that $a_1u_1+...+a_ku_k+bv=0$, then $a_1=...=a_k=b=0$. So $v=a_1u_1+...+a_ku_k$ for some $a_1,...,a_k$.
+
+By definition of linear code, we have $v\in C$, contradicting the assumption.
+
+
+
+### Solution
+
+We proceed by applying the lemma 2.
+
+1. $\operatorname{rank}(H) = n - k$ since $H$ is a Vandermonde matrix times a diagonal matrix with no zero entries, so $H$ is invertible.
+
+2. $H G^T = 0$.
+
+note that $\forall$ row $i$ of $H$, $0\leq i\leq n-k-1$, $\forall$ column $j$ of $G^T$, $0\leq j\leq k-1$
+
+So
+
+$$
+\begin{aligned}
+H G^T &= \begin{bmatrix}
+1 & 1 & \cdots & 1\\
+\alpha_1 & \alpha_2 & \cdots & \alpha_n\\
+\alpha_1^2 & \alpha_2^2 & \cdots & \alpha_n^2\\
+\vdots & \vdots & \cdots & \vdots\\
+\alpha_1^{k-1} & \alpha_2^{k-1} & \cdots & \alpha_n^{k-1}
+\end{bmatrix}\begin{bmatrix}
+v_1 & 0 & \cdots & 0\\
+0 & v_2 & \cdots & 0\\
+\vdots & \vdots & \ddots & \vdots\\
+0 & 0 & \cdots & v_n
+\end{bmatrix}\begin{bmatrix}
+1 & \alpha_1 & \alpha_1^2 & \cdots & \alpha_1^{k-1}\\
+1 & \alpha_2 & \alpha_2^2 & \cdots & \alpha_2^{k-1}\\
+\vdots & \vdots & \vdots & \cdots & \vdots\\
+1 & \alpha_n & \alpha_n^2 & \cdots & \alpha_n^{k-1}
+\end{bmatrix}\\
+&=
+\begin{bmatrix}
+1 & 1 & \cdots & 1\\
+\alpha_1 & \alpha_2 & \cdots & \alpha_n\\
+\alpha_1^2 & \alpha_2^2 & \cdots & \alpha_n^2\\
+\vdots & \vdots & \cdots & \vdots\\
+\alpha_1^{k-1} & \alpha_2^{k-1} & \cdots & \alpha_n^{k-1}
+\end{bmatrix}
+\begin{bmatrix}
+v_1\\
+v_2\\
+\vdots\\
+v_n
+\end{bmatrix}\\
+&=\sum_{l=1}^n\alpha_l^{r}v_l=0
+\end{aligned}
+$$
+
+## Question 3
+
+Show that in an MDS code $[n,k,d]_{\mathbb{F}_q}$, $d=n-k+1$, every $k$ entries determine the remaining $n-k$ entries of $G$.
+
+That is, every $k\times k$ submatrix of $G$ is invertible.
+
+
+Proof
+
+Let $G$ be the generator matrix, and $G'$ be any $k\times k$ submatrix of $G$.
+
+$$
+G=\begin{bmatrix}
+G'\in \mathbb{F}^{k\times k}|G''\in \mathbb{F}^{k\times (n-k)}
+\end{bmatrix}
+$$
+
+We proceed by contradiction, suppose $G'$ is not invertible.
+
+Then there exists $m',m''\in \mathbb{F}_q^k$ such that $m'\neq m''$ but $m'G'=m''G'$.
+
+Note that $m'G=[m'G'|m'G'']$ and $m''G=[m''G'|m''G'']$.
+
+So there are only $n-k$ entries of $m'G$ and $m''G$ are different, so $d\leq n-k$.
+
+That violate with the assumption that $d=n-k+1$.
+
+
+## Reed-Muller code
+
+### Definition of Reed-Muller code (binary case)
+
+$$
+RM(r,m)=\left\{(f(\alpha_1),\ldots,f(\alpha_2^m))|\alpha_i\in \mathbb{F}_2^m,\deg f\leq r\right\}
+$$
+
+Length of $RM(r,m)$ is $2^m$.
+
+
+Example of Reed-Muller code
+
+Let $r=2$, $m=3$.
+
+$\alpha_1=(0,0,0)$, $\alpha_2=(0,0,1)$, $\alpha_3=(0,1,0)$, $\alpha_4=(0,1,1)$, $\alpha_5=(1,0,0)$, $\alpha_6=(1,0,1)$, $\alpha_7=(1,1,0)$, $\alpha_8=(1,1,1)$.
+
+$p(x)\deg\leq 2=\{1,x_1,x_2,x_3,x_1x_2,x_1x_3,x_2x_3\}$.
+
+So $p(x)=1+x_1+x_2+x_3+x_1x_2+x_1x_3+x_2x_3$.
+
+The generator matrix is defined by
+
+$$
+G=\begin{bmatrix}1\\x_1\\x_2\\x_3\\x_1x_2\\x_1x_3\\x_2x_3\end{bmatrix}\begin{bmatrix}
+1& 1&1&1&1&1&1&1\\
+0& 0&0&0&1&1&1&1\\
+0& 0&1&1&0&0&1&1\\
+0& 1&0&1&0&1&0&1\\
+0& 0&0&0&0&0&1&1\\
+0& 0&0&0&0&1&0&1\\
+0& 0&0&1&0&0&0&1\\
+\end{bmatrix}
+$$
+
+
+
+So $\dim RM(r,m)=\sum_{i=0}^{r}\binom{m}{i}$.
+
+## Question 4
+
+$RM(m-1,m)$ is the parity check code.
+
+
+Proof
+
+By previous lemma, it is sufficient to show that $\dim (RM(m-1,m))=n-1$ and $RM(m-1,m)\subseteq \text{ parity code}$.
+
+For the first property,
+
+$$
+\dim (RM(m-1,m))=\sum_{i=0}^{m-1}\binom{m}{i}=\sum_{i=0}^{m}\binom{m}{i}-\binom{m}{m}=2^m-1=n-1
+$$
+
+For the second property,
+
+recall that $c=(c_1,c_2,\ldots,c_n)\in$ parity if $\sum_{i=1}^n c_i=0$.
+
+So we need to show that $\sum_{i=1}^{2^m}f(\alpha_i)=0$ for every $f$ with $\deg f\leq m-1$.
+
+Note that $\forall f\in RM(m-1,m)$, we can write $f(x_1,x_2,\ldots,x_m)=\sum_{S\subseteq [m],|S|\leq m-1}f_Sx_S$
+
+So $\sum_{i=1}^{2^m}f(\alpha_i)=\sum_{S\subseteq [m],|S|\leq m-1}f_S\sum_{i=1}^{2^m}x_S(\alpha_i)=0$.
+
+We denote $J(s)=\sum_{i=1}^{2^m}x_S(\alpha_i)$.
+
+Note that $J(S)=\begin{cases}
+1 & \text{if number of 1 in $S$ is odd}
+0 & \text{otherwise}
+\end{cases}$
+
+And number of 1 in $S$ is $2^{m-|S|}$ which is even.
+
+So $J(S)=0$ for all $S\subseteq [m],|S|\leq m-1$.
+
\ No newline at end of file
diff --git a/content/CSE5313/_meta.js b/content/CSE5313/_meta.js
index 35560ed..3ca12dc 100644
--- a/content/CSE5313/_meta.js
+++ b/content/CSE5313/_meta.js
@@ -12,4 +12,5 @@ export default {
CSE5313_L7: "CSE5313 Coding and information theory for data science (Lecture 7)",
CSE5313_L8: "CSE5313 Coding and information theory for data science (Lecture 8)",
CSE5313_L9: "CSE5313 Coding and information theory for data science (Lecture 9)",
+ CSE5313_L10: "CSE5313 Coding and information theory for data science (Recitation 10)",
}
\ No newline at end of file