grizzly/deneir
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

better formating

Browse source
This commit is contained in:
Arthur Grisel-Davy 2023-06-29 21:32:41 -04:00
parent 5aa4f1169a
commit bd43a83046
4 changed files with 220 additions and 10 deletions
Download patch file Download diff file Expand all files Collapse all files

4
DSD/qrs/biblio.bib
View file

@ -95,7 +95,7 @@ series = {MobiSys '17}
@misc{palitronica,
title = {Palitronica - Palisade},
howpublished = {\url{https://www.palitronica.com/products/palisade}},
howpublished = {https://www.palitronica.com/products/palisade},
note = {Accessed: 2010-03-26}
}
@ -461,7 +461,7 @@ series = {PAMCO '16}
}
@misc{zenodo,
title={Evaluation Dataset for the Machine State Detector, \url{https://zenodo.org/record/7782702#.ZCR33byZNhE}},
title={Evaluation Dataset for the Machine State Detector, https://zenodo.org/record/7782702\#.ZCR33byZNhE},
year={2023},
}

17
DSD/qrs/main.tex
View file

@ -1,5 +1,7 @@
\documentclass[conference]{IEEEconf}
%\input epsf
\usepackage{graphicx}
\usepackage{multirow}
@ -66,7 +68,7 @@ We present in this paper a novel time series, one-shot classifier called \gls{ma
% for peerreview papers, inserts a page break and creates the second title.
% Will be ignored for other modes.
\IEEEpeerreviewmaketitle
\agd{reset acronyms}
\section{Introduction}
@ -228,7 +230,7 @@ ID(P_j,P_l) = \min_{i\in[0,N_l-N_j]} nd(P_j,P_l[i:i+N_j])
\end{equation}
represents the smallest distance between $P_j$ and any substring of length $N_j$ from $P_l$ --- with $N_l>N_j$.
If $N_l<N_j$, then $ID(P_j,P_l) = ID(P_l,P_j)$.
In other words, when computing the inter-distance between two patterns, we slide the short pattern along the length of the long one and compute the normalized distance at every position to finally consider only the smallest of these distances as the inter-distance.
When computing the inter-distance between two patterns, we slide the short pattern along the length of the long one and compute the normalized distance at every position to finally consider only the smallest of these distances as the inter-distance.
To fully define the threshold $T_j$, we introduce the shrinkage coefficient $\alpha$.
This coefficient, multiplied with the smallest inter-distance $ID_j$, forms the threshold $T_j$.
@ -258,7 +260,7 @@ The algorithm for \gls{mad} follows three steps:
However, directly implementing this suite of operations is not optimal as it requires computing the distance from any substring to any pattern multiple times --- exactly once per sample in the substring.
A more efficient solution considers each substring only once.
In other words, iterating over the patterns rather than the samples is more efficient as it replaces distance computations with comparison operations.
Iterating over the patterns rather than the samples is more efficient as it replaces distance computations with comparison operations.
The efficient implementation follows the operations:
\begin{enumerate}
@ -353,8 +355,8 @@ Finally, the third part uses the sames loops as the second and also terminates.
Overall, \gls{mad} always terminates for any finite time series and finite set of finite patterns.
\textbf{Monotony of number of unknown sample}\agd{find better title}
Explain that the number of unknown sample is monotonic as a function of alpha.
Also, a sample that is classified as unknown will always remain unknown if alpha decreases.
\agd{Explain that the number of unknown sample is monotonic as a function of alpha.
Also, a sample that is classified as unknown will always remain unknown if alpha decreases.}
\section{Evaluation}
The evaluation of \gls{mad} consists in the detection of the states for time series from various machines.
@ -452,7 +454,8 @@ The activity in this second time series was very sparse with long periods withou
The no consumption sections: are not a challenging --- i.e., all detectors perform well on this type of pattern ---, make the manual labeling more difficult, and level all results up.
For this reason we removed large sections of inactivity between active segments to make the time series more challenging without tempering with the order of detector performances.
\input{refit_table}
%\input{refit_table}
\agd{include table about refit dataset}
\subsection{Alternative Methods}
\agd{Explain better why the alternative methods are chosen.}
@ -621,7 +624,7 @@ Built as a variation of a traditional \gls{1nn}, \gls{mad} uses a dynamic window
One hyper-parameter, $\alpha$, controls the confidence of the detector and the trade-off between un-classified and miss-classified samples.
The comparison to traditional state detection methods highlights the potential of \gls{mad} for the pre-processing of raw data for security applications.
\bibliographystyle{splncs04}
\bibliographystyle{plain}
\bibliography{biblio}
\end{document}

2
DSD/qrs/refit_table.tex
View file

@ -16,4 +16,4 @@
& & \\
& &
\end{tabular}
\end{table}
\end{table}

207
DSD/qrs/tabto.sty Normal file
View file

@ -0,0 +1,207 @@
% tabto.sty
%
% version 1.4 (Dec 2018)
%
% Tabbing to fixed positions in a paragraph.
%
% Copyright 2006,2009,2012,2013,2018 by
% Donald Arseneau, Vancouver, Canada (asnd@triumf.ca)
% Permission to use, distribute and modify this software is granted
% under the conditions of the LaTeX Project Public License, either
% version 1.3 or (at your option) any later version. The license is
% found at http://www.latex-project.org/lppl.txt, and is part of all
% recent distributions of LaTeX.
%
% This work has the LPPL maintenance status `maintained' (by author).
%
% Two new text positioning commands are defined: \tabto and \tab.
%
% \tabto{<length>}
% Tab to a position relative to the left margin in a paragraph (any
% indentation due to a list or \leftskip is part of the `margin' in
% this context). If the text on the line already goes past the desired
% position, the tab starts a new line and moves to the requested
% horizontal position.
%
% \tabto*{<length>}
% Similar to \tabto, except it will perform backspacing, and over-
% print previous text on the line whenever that text is already
% longer than the specified length (i.e., no linebreak is produced).
% Line-breaks are suppressed immediately after \tabto or \tabto*.
%
% The length register "\CurrentLineWidth" will report the width
% of the existing text on the line, and it may be used in the
% <length> argument (using calc.sty, for example). Also, there
% is "\TabPrevPos" which gives the "\CurrentLineWidth" from the
% previous tab command (the position where the tab command occurred,
% not where it went to), and can be used to return to that position
% if no line breaks have occurred in between, or directly below it,
% if there were line breaks.
%
% \tab
% Tab to the next tab-stop chosen from a list of tab positions, in
% the traditional style of typewriters. A \tab will always move
% to the next tab stop (or the next line), even if it is already
% exactly at a tab stop. Thus, "\tab" at the beginning of a line,
% or "\tab\tab" elsewhere skips a position. A linebreak is permitted
% immediately following a \tab, in case the ensuing text does not
% fit well in the remaining space.
%
% If you do not want to skip positions, use "\tabto{\NextTabStop}"
% instead of "\tab". This is particularly useful when you want to
% use \tab in some other command, but do not want to skip a column
% for the first item.
%
% The tab-stop positions are declared using either \TabPositions
% or \NumTabs:
%
% \TabPositions{<length>, <length>,...<length>}
% Declares the tab stops as a comma-separated list of positions
% relative to the left margin. A tab-stop at 0pt is implicit, and
% need not be listed.
%
% \NumTabs{<number>}
% Declares a list of <number> equally-spaced tabs, starting at the
% left margin and spanning \linewidth. For example \NumTabs{2}
% declares tab-stops at 0pt and 0.5\linewidth, the same as
% \TabPositions{0pt, 0.5\linewidth} or \TabPositions{0.5\linewidth}
%
% After these declarations, the list of tab positions is saved in
% \TabStopList, and the next tab position, relative to the current
% position, is given by \NextTabStop. You do not normally need
% to access them, but they are available.
%
% Problems:
%
% Tall objects after a tab stop may overlap the line above, rather
% than forcing a greater separation between lines.
\ProvidesPackage{tabto}[2018/12/28 \space v 1.4 \space
Another tabbing mechanism]\relax
\newdimen\CurrentLineWidth
\newdimen\TabPrevPos
\newcommand\tabto[1]{%
\leavevmode
\begingroup
\def\@tempa{*}\def\@tempb{#1}%
\ifx\@tempa\@tempb % \tab*
\endgroup
\TTo@overlaptrue % ... set a flag and re-issue \tabto to get argument
\expandafter\tabto
\else
\ifinner % in a \hbox, so ignore
\else % unrestricted horizontal mode
\null% \predisplaysize will tell the position of this box (must be box)
\parfillskip\fill
\everydisplay{}\everymath{}%
\predisplaypenalty\@M \postdisplaypenalty\@M
$$% math display so we can test \predisplaysize
\lineskiplimit=-999pt % so we get pure \baselineskip
\abovedisplayskip=-\baselineskip \abovedisplayshortskip=-\baselineskip
\belowdisplayskip\z@skip \belowdisplayshortskip\z@skip
\halign{##\cr\noalign{%
% get the width of the line above
\ifdim\predisplaysize=\maxdimen %\message{Mixed R and L, so say the line is full. }%
\CurrentLineWidth\linewidth
\else
\ifdim\predisplaysize=-\maxdimen
% \message{Not in a paragraph, so call the line empty. }%
\CurrentLineWidth\z@
\else
\ifnum\TTo@Direction<\z@
\CurrentLineWidth\linewidth \advance\CurrentLineWidth\predisplaysize
\else
\CurrentLineWidth\predisplaysize
\fi
% Correct the 2em offset
\advance\CurrentLineWidth -2em
\advance\CurrentLineWidth -\displayindent
\advance\CurrentLineWidth -\leftskip
\fi\fi
\ifdim\CurrentLineWidth<\z@ \CurrentLineWidth\z@\fi
% Enshrine the tab-to position; #1 might reference \CurrentLineWidth
\setlength\@tempdimb{#1}% allow calc.sty
%\message{*** Tab to \the\@tempdimb, previous width is \the\CurrentLineWidth. ***}%
% Save width for possible return use
\global\TabPrevPos\CurrentLineWidth
% Build the action to perform
\protected@xdef\TTo@action{%
\vrule\@width\z@\@depth\the\prevdepth
\ifdim\CurrentLineWidth>\@tempdimb
\ifTTo@overlap\else
\protect\newline \protect\null
\fi\fi
\protect\nobreak
\protect\hskip\the\@tempdimb\relax
}%
%\message{\string\TTo@action: \meaning \TTo@action. }%
% get back to the baseline, regardless of its depth.
\vskip-\prevdepth
\prevdepth-99\p@
\vskip\prevdepth
}}%
$$
% Don't count the display as lines in the paragraph
\count@\prevgraf \advance\count@-4 \prevgraf\count@
\TTo@action
%% \penalty\@m % to allow a penalized line break
\fi
\endgroup
\TTo@overlapfalse
\ignorespaces
\fi
}
% \tab -- to the next position
% \hskip so \tab\tab moves two positions
% Allow a (penalized but flexible) line-break right after the tab.
%
\newcommand\tab{\leavevmode\hskip2sp\tabto{\NextTabStop}%
\nobreak\hskip\z@\@plus 30\p@\penalty4000\hskip\z@\@plus-30\p@\relax}
% Expandable macro to select the next tab position from the list
\newcommand\NextTabStop{%
\expandafter \TTo@nexttabstop \TabStopList,\maxdimen,>%
}
\def\TTo@nexttabstop #1,{%
\ifdim#1<\CurrentLineWidth
\expandafter\TTo@nexttabstop
\else
\ifdim#1<0.9999\linewidth#1\else\z@\fi
\expandafter\strip@prefix
\fi
}
\def\TTo@foundtabstop#1>{}
\newcommand\TabPositions[1]{\def\TabStopList{\z@,#1}}
\newcommand\NumTabs[1]{%
\def\TabStopList{}%
\@tempdimb\linewidth
\divide\@tempdimb by#1\relax
\advance\@tempdimb 1sp % counteract rounding-down by \divide
\CurrentLineWidth\z@
\@whiledim\CurrentLineWidth<\linewidth\do {%
\edef\TabStopList{\TabStopList\the\CurrentLineWidth,}%
\advance\CurrentLineWidth\@tempdimb
}%
\edef\TabStopList{\TabStopList\linewidth}%
}
% default setting of tab positions:
\TabPositions{\parindent,.5\linewidth}
\newif\ifTTo@overlap \TTo@overlapfalse
\@ifundefined{predisplaydirection}{
\let\TTo@Direction\predisplaysize
\let\predisplaydirection\@undefined
}{
\let\TTo@Direction\predisplaydirection
}