文章目录:
-
基本论文结构
-
算法
-
图片
-
表格
-
公式
-
特殊符号
-
参考文献
-
序号
-
总结
基本论文结构
什么是LaTeX
基本结构
% 导言区
documentclass{article}
title{My First Document}
author{Eastmount}
date{today}
% 正文区
begin{document}
maketitle
Hello World!
end{document}
篇章结构
导言区
documentclass{article}
usepackage{ctex}
title{My First Document}
author{Eastmount}
date{today}
% 正文区
begin{document}
maketitle
构建文章小结
section{Introduction}
section{Related Work}
section{System Model}
section{Mathematics and algorithms}
section{Experiments}
subsection{Datasets}
subsubsection{实验条件}
subsubsection{评价指标}
subsection{Results}
section{Acknowledgment}
end{document}
算法
algorithm
usepackage{algorithm}
usepackage{algorithmic}
begin{algorithm}[!ht]
caption{Feature extraction based on abstract syntax tree.}
begin{algorithmic}[1]
REQUIRE {$X=left{x_1,x_2,...,x_nright}$, where $x_i$ is the $i^{th}$ PowerShell script.}
ENSURE {$V^{(ast)}=left{v_1,v_2,...,v_nright}$, where $v_i$ is the $i^{th}$ sequence vector generated by AST-based feature extraction method (i.e., AST2Vec).}
STATE $V^{(ast)} leftarrow emptyset$ , $S leftarrow emptyset$, $F leftarrow emptyset$, $W leftarrow emptyset$
FOR{$i leftarrow 1$ {bf to} $n$}
STATE $t_i = ExtractAstSequences( x_i )$
STATE $s_i = PostorderTraversal( t_i )$
STATE $S.append( s_i )$
ENDFOR
STATE $F = BuildFeatureSetFromAst(S) $
STATE $//$ Count all distinct features of AST sequences.
FOR{each $f_k in F$}
STATE $w_k = CalculateWordVectors( f_k )$
STATE $W.append( w_k )$
ENDFOR
STATE $//$ Calculate word vectors for each AST node.
FOR{each $s_i in S$}
STATE $v_i = GenerateAstEmbedding( s_i )$
STATE $V^{(ast)}.append( v_i )$
ENDFOR
STATE {bf return} {$V^{(ast)}$}
end{algorithmic}
label{algorithm1}
end{algorithm}
algorithm2e
usepackage{algorithm}
usepackage[algo2e]{algorithm2e}
begin{algorithm}[!ht]
caption{Feature extraction based on abstract syntax tree.}
label{algorithm1}
SetAlgoLined
SetKwInOut{Input}{Input}
SetKwInOut{Output}{Output}
Input{$X=left{x_1,x_2,...,x_nright}$, where $x_i$ is the $i^{th}$ PowerShell script.}
Output{$V^{(ast)}=left{v_1,v_2,...,v_nright}$, where $v_i$ is the $i^{th}$ sequence vector generated by AST-based feature extraction method (i.e., AST2Vec).}
Initialization: $V^{(ast)} leftarrow emptyset$ , $S leftarrow emptyset$, $F leftarrow emptyset$, $W leftarrow emptyset$
For{$i leftarrow 1$ KwTo $n$}{
$t_i = ExtractAstSequences( x_i )$
$s_i = PostorderTraversal( t_i )$
$S.append( s_i )$
}
$F = BuildFeatureSetFromAst(S) $
tcc{Count all distinct features of AST sequences}
For{$f_k in F$}{
$w_k = CalculateWordVectors( f_k )$
$W.append( w_k )$
}
tcc{Calculate word vectors for each AST node}
For{$s_i in S$}{
$v_i = GenerateAstEmbedding( s_i )$
$V^{(ast)}.append( v_i )$
}
Return{$V^{(ast)}$}
end{algorithm}
图片
基本用法
导言区
documentclass{article}
usepackage{ctex}
usepackage{graphicx}
% 指定图片在当前目录下figures目录下
graphicspath{{figures/}}
% 正文区
begin{document}
插入图片
includegraphics{fig1}
缩放比例
includegraphics[scale=0.5]{fig1}
固定图像高度
includegraphics[height=2cm]{fig1.png}
固定图像宽度
includegraphics[width=2cm]{fig1.png}
图像高度和宽度基于
includegraphics[height=0.2textheight]{fig1.png}
includegraphics[width=0.2textwidth]{fig1.png}
指定多个参数
includegraphics[angle=-45,width=0.5textwidth]{fig1.png}
end{document}
双栏显示
usepackage{stfloats}
begin{figure*}[ht]
centering
includegraphics[width=0.8textwidth]{4.eps}
caption{BER performance of original OFDM system and different companding schemes over AWGN channel (QPSK).}
label{fig8}
end{figure*}
h——放置在此处
t——放置在顶部
b——放置在底部
p——浮动放置
*——两栏放置
双图显示
begin{figure*}
begin{minipage}[t]{0.48linewidth}
centering
includegraphics[scale=0.30]{Figure-7.eps}
caption{The loss curve of different models.}
label{fig7}
end{minipage}
begin{minipage}[t]{0.48linewidth}
centering
includegraphics[scale=0.30]{Figure-8.eps}
caption{The accuracy curve of different models.}
label{fig8}
end{minipage}
end{figure*}
usepackage{caption}
usepackage{subfigure}
begin{figure}[htbp]
centering %居中
subfigure[name of the first figure] %第一张子图
{
begin{minipage}[t]{0.4textwidth}
centering
includegraphics[scale=0.15]{fig2}
end{minipage}
}
subfigure[name of the second figure] %第二张子图
{
begin{minipage}[t]{0.4textwidth}
centering
includegraphics[scale=0.2]{fig3}
end{minipage}
}
caption{name of the figure} %大图名称
label{fig-1} %图片引用标记
end{figure}
begin{figure}[!hb]
centering
subfloat[label{fig:arm1}$Q^{*}$ values for arm 1]{includegraphics[width=.5linewidth]{1.eps}}%
subfloat[label{fig:arm2}$Q^{*}$ values for arm 2]{includegraphics[width=.5linewidth]{1.eps}}\
subfloat[label{fig:arm3}$Q^{*}$ values for arm 3]{includegraphics[width=.8linewidth]{1.eps}}
caption{$Q^{*}$ values for different arms.}
label{fig:arms}
end{figure}
表格
基本用法
begin{table}
caption{Symbol Table}
centering
begin{tabular}{lll}
hline
Symbol & Definition & Unitis\
noalign{globalarrayrulewidth1pt}hlinenoalign{globalarrayrulewidth0.4pt}
multicolumn{3}{c}{textbf{Constants}}\
$lambda$ & Mean of Poisson distribution & unitless\
$p_{slow}$ & Probability that a vehicle slows down randomly & unitless\
hline
end{tabular}
end{table}
含注释的表格
usepackage{threeparttable}
usepackage{float}
usepackage{bbding}
usepackage{pifont}
begin{table*}[!ht]
centering
caption{Related work comparison.}
begin{threeparttable}
resizebox{textwidth}{!}{
begin{tabular}{ccccm{1.2cm}<{centering}m{1.2cm}<{centering}cc}hline
Related work & Techniques & Focus & Deobfuscation & AST & KG & Multi-modal & Transformer \hline
Li et al. cite{b3} & makecell[c]{subtree-based deobfuscation \ OOA mining algorithm} & deobfuscation & Checkmark & Checkmark & XSolid & XSolid & XSolid \hline
PSDEM cite{b12} & makecell[c]{two-layer deobfuscation \ monitor process by dynamic analysis} & deobfuscation & Checkmark & XSolid & XSolid & XSolid & XSolid \hline
PowerDrive cite{b13} & makecell[c]{multi-stage deobfuscator \ static analysis by regex \ dynamic analysis by cmdlet} & deobfuscation & Checkmark & XSolid & XSolid & XSolid & XSolid \hline
PowerDecode cite{b15} & makecell[c]{syntax check and remove base64 encoding \ deobfuscate by cmdlet overriding \ deobfuscate by regex} & deobfuscation & Checkmark & XSolid & XSolid & XSolid & XSolid \hline
Hendler et al. cite{b18} & makecell[c]{character-level CNN \ 4-layer CNN} & binary classification & XSolid & XSolid & XSolid & XSolid & XSolid \hline
Fang et al. cite{b19} & makecell[c]{hybrid features \ FastText \ random forest} & binary classification & Checkmark & Checkmark & XSolid & $bigcirc$ & XSolid \hline
Hendler et al. cite{b2} & makecell[c]{AMSI-based detection \ contextual embeddings \ Token-Char-FastText} & binary classification & XSolid & XSolid & XSolid & XSolid & XSolid \hline
FC-PSDS cite{b25} & makecell[c]{ features combination \ random forest and DNN} & binary classification & Checkmark & Checkmark & XSolid & XSolid & XSolid \hline
Ruscak et al. cite{b20} & makecell[c]{abstract syntax tree \ random forest} & multi-classification task & XSolid & Checkmark & XSolid & XSolid & XSolid \hline
makecell[c]{textbf{Our method} \ textbf{PowerDetector}} & makecell[c]{multi-modal embedding \ Transformer-CNN-BiLSTM \ multi-layer deobfuscation algorithm } & makecell[c]{malicious family detection \ multi-classification task} & Checkmark & Checkmark & Checkmark & Checkmark & Checkmark \hline
end{tabular}}
begin{tablenotes}
footnotesize
item In this table, Checkmark stands for fully cover, $bigcirc$ stands for partial cover, XSolid means cannot cover.
end{tablenotes}
end{threeparttable}
label{tab1}
end{table*}
复杂表格合并multirow
usepackage{multirow}
begin{table*}[!ht]
centering
caption{Detailed performance comparison of single-modal and multi-modal.}
resizebox{textwidth}{!}{
begin{tabular}{ccccccccccc}hline
multirow{3}{*}{Model} & multicolumn{8}{c}{Single-modal} & multicolumn{2}{c}{multirow{2}{*}{Multi-modal}} \
cline{2-9}
& multicolumn{2}{c}{Token-level} & multicolumn{2}{c}{Character-level} & multicolumn{2}{c}{AST-level} &
multicolumn{2}{c}{KG-level} & multicolumn{2}{c}{} \
cline{2-11}
& $F_1$ & Acc & $F_1$ & Acc & $F_1$ & Acc & $F_1$ & Acc & $F_1$ & Acc \hline
LR & 0.8727 & 0.8629 & 0.8496 & 0.8528 & 0.8661 & 0.8700 & 0.8646 & 0.8559 & 0.8895 & 0.8857 \
RF & 0.8723 & textbf{0.8676} & textbf{0.8610} & textbf{0.8567} & textbf{0.8807} & textbf{0.8786} & 0.8723 & 0.8676 & textbf{0.9017} & textbf{0.8943} \
SVM & 0.8764 & 0.8661 & 0.8527 & 0.8519 & 0.8755 & 0.8786 & textbf{0.8771} & 0.8676 & 0.8934 & 0.8912 \
KNN & 0.8706 & 0.8669 & 0.8554 & 0.8536 & 0.8644 & 0.8637 & 0.8741 & textbf{0.8715} & 0.8804 & 0.8771 \hline
CNN & 0.9002 & 0.8974 & 0.8826 & 0.8808 & 0.9019 & 0.8998 & 0.9025 & 0.8998 & 0.9153 & 0.9115 \
TextCNN & 0.9049 & 0.9013 & 0.9012 & 0.8966 & 0.9083 & 0.9076 & 0.9036 & 0.9005 & 0.9186 & 0.9178 \
BiLSTM & 0.9076 & 0.9069 & 0.9037 & 0.9036 & 0.9126 & 0.9107 & 0.9054 & 0.9025 & 0.9226 & 0.9209 \
BiGRU & 0.9041 & 0.9021 & 0.8989 & 0.8966 & 0.9092 & 0.9045 & 0.9046 & 0.9013 & 0.9205 & 0.9201 \
Transformer & 0.9123 & 0.9107 & 0.9053 & 0.9029 & 0.9116 & 0.9092 & 0.9121 & 0.9115 & 0.9224 & 0.9178 \
CNN-BiLSTM+ATT & textbf{0.9142} & textbf{0.9123} & textbf{0.9081} & textbf{0.9076} & textbf{0.9144} & textbf{0.9139} & textbf{0.9139} & textbf{0.9123} & textbf{0.9262} & textbf{0.9209} \hline
textbf{Our Method} & textbf{0.9236} & textbf{0.9225} & textbf{0.9170} & textbf{0.9169} & textbf{0.9248} & textbf{0.9233} & textbf{0.9204} & textbf{0.9201} & textbf{0.9374} &textbf{0.9358} \hline
end{tabular}}
label{tab5}
end{table*}
解决自动换行
begin{table*}[h]
centering
begin{tabular}{cc} hline
Use Case Navn: & Opret Server \hline
Scenarie: & At oprette en server med bestemte regler som tillader folk at spille sammen.
The nonlinear companding function introduce some nonlinear distortion to original OFDM signal,
which can be eliminated theoretically by the decompanding function. \hline
end{tabular}
label{tab5}
end{table*}
usepackage{tabularx}
begin{table*}[h]
centering
begin{tabularx}{textwidth}{p{4cm} X} hline
Use Case Navn: & Opret Server \hline
Scenarie: & At oprette en server med bestemte regler som tillader folk at spille sammen.
The nonlinear companding function introduce some nonlinear distortion to original OFDM signal,
which can be eliminated theoretically by the decompanding function. \hline
end{tabularx}
label{tab5}
end{table*}
公式
begin{equation}
begin{aligned}
V^{(token)}=
begin{bmatrix}
v_{11} & v_{12} & cdots & v_{1m} \
v_{21} & v_{22} & cdots & v_{2m} \
vdots & vdots & ddots & vdots \
v_{n1} & v_{n2} & cdots & v_{nm} \
end{bmatrix}
end{aligned}
label{eq1}
end{equation}
begin{equation}
begin{aligned}
TokenPairs[j][k] =
begin{cases}
v_{jk} + 1 & exists <f_j,f_k> \
v_{jk} + 0 & other \
end{cases}.
end{aligned}label{eq2}
end{equation}
begin{equation}
begin{aligned}
Attention(textbf{Q},textbf{K},textbf{V}) = softmax left( frac{textbf{Q}textbf{K}^T}{sqrt{d_k}} right) textbf{V}.
end{aligned}label{eq5}
end{equation}
begin{equation}
begin{aligned}
Accuracy = sum_{i=1}^N Accuracy_i times w_i .
end{aligned}label{eq15}
end{equation}
特殊符号
圆圈数字
usepackage{pifont}
ding{184}
ding{182}ding{183}ding{184}ding{185}ding{186}ding{187}ding{188}ding{189}ding{190}ding{191}\
ding{192}ding{193}ding{194}ding{195}ding{196}ding{197}ding{198}ding{199}ding{200}ding{201}\
ding{202}ding{203}ding{204}ding{205}ding{206}ding{207}ding{208}ding{209}ding{210}ding{211}\
textcircled{3}$
normalsize{textcircled{scriptsize{3}}}normalsizeenspace
半圆
usepackage{tikz}
newcommand*emptycirc[1][1ex]{tikzdraw (0,0) circle (#1);}
newcommand*halfcirc[1][1ex]{%
begin{tikzpicture}
draw[fill] (0,0)-- (90:#1) arc (90:270:#1) -- cycle ;
draw (0,0) circle (#1);
end{tikzpicture}}
newcommand*fullcirc[1][1ex]{tikzfill (0,0) circle (#1);}
fullcirc
halfcirc
emptycirc
勾叉
usepackage{pifont} % ding{xx}
usepackage{bbding} % Checkmark,XSolid,... (需要和pifont宏包共同使用)
checkmark
Checkmark
CheckmarkBold
XSolid
XSolidBold
XSolidBrush
纸牌
$clubsuit$
$spadesuit$
$heartsuit$
$diamondsuit$
参考文献
begin{thebibliography}{99}
bibitem{ref1}Zheng L, Wang S, Tian L, et al., Query-adaptive late fusion for image search and person re-identification, Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2015: 1741-1750.
bibitem{ref2}Arandjelović R, Zisserman A, Three things everyone should know to improve object retrieval, Computer Vision and Pattern Recognition (CVPR), 2012 IEEE Conference on, IEEE, 2012: 2911-2918.
bibitem{ref3}Lowe D G. Distinctive image features from scale-invariant keypoints, International journal of computer vision, 2004, 60(2): 91-110.
bibitem{ref4}Philbin J, Chum O, Isard M, et al. Lost in quantization: Improving particular object retrieval in large scale image databases, Computer Vision and Pattern Recognition, 2008. CVPR 2008, IEEE Conference on, IEEE, 2008: 1-8.
end{thebibliography}
begin{thebibliography}{49}
bibitem{b1} Microsoft, ``What is PowerShell? - PowerShell | Microsoft Docs,''
Website: https://docs.microsoft.com/en-us/powershell/scripting/overview, 2022.
bibitem{b2} D. Hendler, S. Kels, et al., ``AMSI-Based Detection of Malicious PowerShell
Code Using Contextual Embeddings,'' in 15th ACM Asia Conference on Computer and
Communications Security (AsiaCCS). ACM, 2020, pp. 679-693.
bibitem{b49} M. Ring, D. Schlor, et al., ``Malware detection on windows audit logs using
LSTMs,'' Computers & Security, vol.109, 2021, p. 102389.
end{thebibliography}
@misc{b1,
title = {What is PowerShell? - PowerShell | Microsoft Docs},
url = {https://docs.microsoft.com/en-us/powershell/scripting/overview},
author = {Microsoft},
year = {2022}
}
@inproceedings{b2,
title={Amsi-based detection of malicious powershell code using contextual embeddings},
author={Hendler, Danny and Kels, Shay and Rubin, Amir},
booktitle={Proceedings of the 15th ACM Asia Conference on Computer and Communications Security (AsiaCCS)},
pages={679--693},
year={2020},
organization = {ACM}
}
@article{b49,
title={Malware detection on windows audit logs using LSTMs},
author={Ring, Markus and Schl{"o}r, Daniel and Wunderlich, Sarah and Landes, Dieter and Hotho, Andreas},
journal={Computers & Security},
volume={109},
pages={102389},
year={2021},
publisher={Elsevier}
}
序号
begin{itemize}
item Every sentence should make sense.
item There is a lot to be said.
item Eschew the highfalutin.
end{itemize}
begin{enumerate}
item Every sentence should make sense.
item There is a lot to be said.
item Eschew the highfalutin.
end{enumerate}
begin{description}
item[Rule 1.] Every sentence should make sense.
item[Rule 2.] There is a lot to be said.
item[Rule 3.] Eschew the highfalutin.
end{description}
评论前必须登录!
立即登录 注册