From 489b32d35ab6bf637f1c471c355c8eb5d66b153e Mon Sep 17 00:00:00 2001 From: Joaquim Castilla Date: Thu, 24 Sep 2020 01:51:36 +0900 Subject: [PATCH] [TR] 13-3: propagating fix to English latex made in https://github.com/Atcold/pytorch-Deep-Learning/pull/576 (#617) --- docs/tr/week13/13-3.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/tr/week13/13-3.md b/docs/tr/week13/13-3.md index 7a2b91e78..cb4cee15f 100644 --- a/docs/tr/week13/13-3.md +++ b/docs/tr/week13/13-3.md @@ -44,7 +44,7 @@ Unlike a sequence, it does not have an order. Fig. 1: Graph Convolutional Network In Figure 1, vertex $v$ is comprised of two vectors: input $\boldsymbol{x}$ and its hidden representation $\boldsymbol{h}$. -We also have multiple vertices $v_{j}$, which is comprised of $\boldsymbol{x}_{j}$ and $\boldsymbol{h}_{j}$. +We also have multiple vertices $v_{j}$, which is comprised of $\boldsymbol{x}\_j$ and $\boldsymbol{h}\_j$. In this graph, vertices are connected with directed edges. We represent these directed edges with adjacency vector $\boldsymbol{a}$, where each element $\alpha_{j}$ is set to $1$ if there is a directed edge from $v_{j}$ to $v$. $$ @@ -82,7 +82,7 @@ where $\vect{D} = \text{diag}(d_{i})$. Şekil 1'de, düğüm *(vertex)* $v$ iki vektörden oluşur: girdi $\boldsymbol{x}$ ve saklı gösterimi $\boldsymbol{h}$. -Çizgemizde $v_{j}$, which is comprised of $\boldsymbol{x}_{j}$ ,\text{and} $\boldsymbol{h}_{j}$. Üstteki çizgede düğümler, yönlü ayrıtlar *(directed edges)* ile birbirlerine bağlıdır. +Çizgemizde $v_{j}$, which is comprised of $\boldsymbol{x}\_j$ and $\boldsymbol{h}\_j$. Üstteki çizgede düğümler, yönlü ayrıtlar *(directed edges)* ile birbirlerine bağlıdır. Yönlü ayrıtlar, komşuluk vektörü $\boldsymbol{a}$ ile ifade edilir, bu vektörün elemanı $\alpha_{j}$ eğer $v_{j}$'den $v$'ye gidilebiliyorsa 1 değerini alır.