mirror of
https://github.com/MartinThoma/LaTeX-examples.git
synced 2025-04-26 06:48:04 +02:00
Bachelor-Final-Presentation: Added first draft
This commit is contained in:
Martin Thoma
parent
22b5998d56
commit
1ee32917ab
30 changed files with 544 additions and 0 deletions
|
|
@ -0,0 +1,47 @@
|
|||
\documentclass[hyperref={pdfpagelabels=false},usepdftitle=false]{beamer}
|
||||
\usepackage{/home/moose/Downloads/LaTeX-examples/presentations/Bachelor-Final-Presentation/templates/myStyle}
|
||||
|
||||
\begin{document}
|
||||
\selectlanguage{english}
|
||||
|
||||
\title{\titleText}
|
||||
\subtitle{Bachelor's thesis of Martin Thoma}
|
||||
\author{\tutor}
|
||||
\date{5th of June, 2014}
|
||||
%\subject{Programmieren}
|
||||
|
||||
\frame{\titlepage}
|
||||
|
||||
\frame{
|
||||
\frametitle{Contents}
|
||||
\setcounter{tocdepth}{1}
|
||||
\tableofcontents
|
||||
\setcounter{tocdepth}{2}
|
||||
}
|
||||
|
||||
%\AtBeginSection[]{
|
||||
% \InsertToC[sections={\thesection}] % shows only subsubsections of one subsection
|
||||
%}
|
||||
|
||||
\section{What is my Bachelor's thesis about?}
|
||||
\input{introduction}
|
||||
|
||||
\section{write-math.com and HWRT}
|
||||
\input{write-math}
|
||||
|
||||
% \section{Preprocessing and Features}
|
||||
% \input{preprocessing}
|
||||
% \input{features}
|
||||
|
||||
\section{Evaluation}
|
||||
\input{evaluation}
|
||||
|
||||
% \section{What will I do next?}
|
||||
% \input{will-do}
|
||||
|
||||
\section*{End}
|
||||
\subsection{End}
|
||||
\input{sources}
|
||||
\framedgraphic{Thanks for Your Attention!}{../images/xi.png}
|
||||
|
||||
\end{document}
|
||||
10
presentations/Bachelor-Final-Presentation/LaTeX/Makefile
Normal file
10
presentations/Bachelor-Final-Presentation/LaTeX/Makefile
Normal file
|
|
@ -0,0 +1,10 @@
|
|||
SOURCE = Bachelor-Final-Presentation
|
||||
|
||||
make:
|
||||
#latexmk -pdf -pdflatex="pdflatex -interactive=nonstopmode" -use-make $(SOURCE).tex
|
||||
pdflatex -shell-escape $(SOURCE).tex -output-format=pdf #shellescape wird fürs logo benötigt
|
||||
pdflatex -shell-escape $(SOURCE).tex -output-format=pdf # nochmaliges ausführen wegen Inhaltsverzeichnissen
|
||||
make clean
|
||||
|
||||
clean:
|
||||
rm -rf $(TARGET) *.class *.html *.log *.aux *.out *.glo *.glg *.gls *.ist *.xdy *.1 *.toc *.snm *.nav *.vrb *.fls *.fdb_latexmk *.pyg
|
||||
111
presentations/Bachelor-Final-Presentation/LaTeX/evaluation.tex
Normal file
111
presentations/Bachelor-Final-Presentation/LaTeX/evaluation.tex
Normal file
|
|
@ -0,0 +1,111 @@
|
|||
\subsection{Merged Symbols}
|
||||
\begin{frame}[fragile]{Merged symbols (MER error)}
|
||||
\begin{table}[ht]
|
||||
\centering
|
||||
\begin{tabular}{lc|lc}
|
||||
\toprule
|
||||
\multicolumn{2}{c}{Base symbol} & \multicolumn{2}{c}{equivalent symbols}\\
|
||||
\LaTeX & Rendered & \LaTeX & Rendered \\\midrule
|
||||
\verb+\sum+ & $\sum$ & \verb+$\Sigma$+ & $\Sigma$\\
|
||||
\verb+\prod+ & $\prod$ & \verb+$\Pi$+ & $\Pi$\\
|
||||
~ & ~ & \verb+$\sqcap$+ & $\sqcap$\\
|
||||
\verb+\coprod+ & $\coprod$ & \verb+$\amalg$+ & $\amalg$\\
|
||||
~ & ~ & \verb+$\sqcup$+ & $\sqcup$\\
|
||||
\verb+\perp+ & $\perp$ & \verb+$\bot$+ & $\bot$\\
|
||||
\verb+\models+ & $\models$ & \verb+$\vDash$+ & $\vDash$\\
|
||||
\verb+|+ & | & \verb+\mid+ & $\mid$ \\
|
||||
\verb+\Delta+ & $\Delta$ & \verb+$\triangle$+ & $\triangle$\\
|
||||
~ & ~ & \verb+$\vartriangle$+ & $\vartriangle$\\
|
||||
\bottomrule
|
||||
\end{tabular}
|
||||
\end{table}
|
||||
\end{frame}
|
||||
|
||||
\begin{frame}[fragile]{Merged symbols (MER error)}
|
||||
\begin{table}[ht]
|
||||
\centering
|
||||
\begin{tabular}{lc|lc}
|
||||
\toprule
|
||||
\multicolumn{2}{c}{Base symbol} & \multicolumn{2}{c}{equivalent symbols}\\
|
||||
\LaTeX & Rendered & \LaTeX & Rendered \\\midrule
|
||||
\verb+\|+ & $\|$ & \verb+$\parallel$+ & $\parallel$\\
|
||||
\verb+\ohm+ & $\Omega$ & \verb+$\Omega$+ & $\Omega$\\
|
||||
\verb+\setminus+ & $\setminus$ & \verb+$\backslash$+ & $\backslash$\\
|
||||
\verb+\checked+ & {\mbox {\wasyfamily \char 8}} & \verb+$\checkmark$+ & $\checkmark$\\
|
||||
\verb+\&+ & $\&$ & \verb+$\with$+ & $\&$\\
|
||||
\verb+\#+ & $\#$ & \verb+$\sharp$+ & $\sharp$\\
|
||||
\verb+\S+ & $\S$ & \verb+$\mathsection$+ & $\mathsection$\\
|
||||
\verb+\nabla+ & $\nabla$ & \verb+\triangledown+ & $\triangledown$\\
|
||||
\verb+\lhd+ & $\lhd$ & \verb+$\triangleleft$+ & $\triangleleft$\\
|
||||
~ & ~ & \verb+$\vartriangleleft$+ & $\vartriangleleft$\\
|
||||
\verb+\oiint+ & $\oiint$ & \verb+$\varoiint$+ & $\varoiint$\\
|
||||
\bottomrule
|
||||
\end{tabular}
|
||||
\end{table}
|
||||
\end{frame}
|
||||
|
||||
|
||||
\begin{frame}[fragile]{Merged symbols (MER error)}
|
||||
\begin{table}[ht]
|
||||
\centering
|
||||
\begin{tabular}{lc|lc}
|
||||
\toprule
|
||||
\multicolumn{2}{c}{Base symbol} & \multicolumn{2}{c}{equivalent symbols}\\
|
||||
\LaTeX & Rendered & \LaTeX & Rendered \\\midrule
|
||||
\verb+\mathbb{R}+ & $\mathbb{R}$ & \verb+$\mathds{R}$+ & $\mathds{R}$\\
|
||||
\verb+\mathbb{Q}+ & $\mathbb{Q}$ & \verb+\mathds{Q}+ & $\mathds{Q}$\\
|
||||
\verb+\mathbb{Z}+ & $\mathbb{Z}$ & \verb+\mathds{Z}+ & $\mathds{Z}$\\
|
||||
\verb+\mathcal{A}+ & $\mathcal{A}$ & \verb+\mathscr{A}+ & $\mathscr{A}$\\
|
||||
\verb+\mathcal{D}+ & $\mathcal{D}$ & \verb+\mathscr{D}+ & $\mathscr{D}$\\
|
||||
\verb+\mathcal{N}+ & $\mathcal{N}$ & \verb+\mathscr{N}+ & $\mathscr{N}$\\
|
||||
\verb+\mathcal{R}+ & $\mathcal{R}$ & \verb+\mathscr{R}+ & $\mathscr{R}$\\
|
||||
\verb+\propto+ & $\propto$ & \verb+$\varpropto$+ & $\varpropto$\\
|
||||
\bottomrule
|
||||
\end{tabular}
|
||||
\end{table}
|
||||
\end{frame}
|
||||
|
||||
\subsection{Evaluation}
|
||||
\begin{frame}{Experiments}
|
||||
\textbf{Preprocessing:} Scaling, shifting and linear interpolation\\
|
||||
\textbf{Features:} Coordinates of 80 points (4 strokes with 20 points each)\\
|
||||
\textbf{Learning:} MLP, 1000 epochs, LR $\eta=0.1$, Momentum $\alpha=0.1$
|
||||
\begin{table}[h]
|
||||
\centering
|
||||
\begin{tabular}{clrrr}
|
||||
\toprule
|
||||
\multirow{2}{*}{System} & \multirow{2}{*}{Topology} & \multicolumn{3}{c}{Classification error}\\
|
||||
& & TOP1 & TOP3 & MER \\\midrule
|
||||
$B_1$ & 160:500:369 & $\SI{23.34}{\percent}$ & $\SI{6.80}{\percent}$ & $\SI{6.64}{\percent}$ \\
|
||||
$B_2$ & 160:500:500:369 & \underline{$\SI{21.51}{\percent}$} & $\SI{5.75}{\percent}$ & $\SI{5.67}{\percent}$ \\
|
||||
$B_3$ & 160:500:500:500:369 & $\SI{21.93}{\percent}$ & \underline{$\SI{5.74}{\percent}$} & \underline{$\SI{5.64}{\percent}$} \\
|
||||
$B_4$ & 160:500:500:500:500:369 & $\SI{23.88}{\percent}$ & $\SI{6.12}{\percent}$ & $\SI{6.04}{\percent}$ \\
|
||||
\bottomrule
|
||||
\end{tabular}
|
||||
\caption{Baseline systems with three different classification error measures.
|
||||
All errors were measured on the test set.}
|
||||
\label{table:baseline-systems}
|
||||
\end{table}
|
||||
\end{frame}
|
||||
|
||||
\subsection{Complex classifier}
|
||||
\begin{frame}{Complex classifier}
|
||||
\textbf{Preprocessing:} Connect strokes, scale, shift and linear interpolation\\
|
||||
\textbf{Features:} Coordinates of 80 points (4 strokes with 20 points each), re-curvature per stroke, ink, stroke count, aspect ratio\\
|
||||
\textbf{Learning:} MLP, 1000 epochs, LR $\eta=0.1$, Momentum $\alpha=0.1$
|
||||
\begin{table}[htb]
|
||||
\centering
|
||||
\begin{tabular}{lrrrrrr}
|
||||
\toprule
|
||||
\multirow{2}{*}{System} & \multicolumn{3}{c}{Classification error}\\
|
||||
& TOP1 & change & TOP3 & change & MER & change \\\midrule
|
||||
$B_{1,c}$ & $\SI{20.96}{\percent}$ & $\SI{-2.38}{\percent}$ & $\SI{5.24}{\percent}$ & $\SI{-1.56}{\percent}$ & $\SI{5.13}{\percent}$ & $\SI{-1.51}{\percent}$ \\
|
||||
$B_{2,c}$ & $\SI{20.10}{\percent}$ & $\SI{-1.41}{\percent}$ & $\SI{4.44}{\percent}$ & $\SI{-1.31}{\percent}$ & $\SI{4.36}{\percent}$ & $\SI{-1.31}{\percent}$ \\
|
||||
$B_{3,c}$ & $\SI{21.51}{\percent}$ & $\SI{-0.42}{\percent}$ & $\SI{4.89}{\percent}$ & $\SI{-0.85}{\percent}$ & $\SI{4.80}{\percent}$ & $\SI{-0.84}{\percent}$ \\
|
||||
$B_{4,c}$ & $\SI{00.00}{\percent}$ & $\SI{-0.00}{\percent}$ & $\SI{0.00}{\percent}$ & $\SI{-0.00}{\percent}$ & $\SI{0.00}{\percent}$ & $\SI{-0.00}{\percent}$ \\
|
||||
\bottomrule
|
||||
\end{tabular}
|
||||
\caption{Error rates of the complex recognizer systems.}
|
||||
\label{table:complex-recognizer-systems-evaluation}
|
||||
\end{table}
|
||||
\end{frame}
|
||||
23
presentations/Bachelor-Final-Presentation/LaTeX/features.tex
Normal file
23
presentations/Bachelor-Final-Presentation/LaTeX/features.tex
Normal file
|
|
@ -0,0 +1,23 @@
|
|||
\subsection{Features}
|
||||
\begin{frame}{Features}
|
||||
\begin{itemize}
|
||||
\item Local
|
||||
\begin{itemize}
|
||||
\item Coordinates
|
||||
\item Speed
|
||||
\item Binary pen pressure
|
||||
\item Direction
|
||||
\item Curvature
|
||||
\item Bitmap-environment
|
||||
\item Hat-Feature
|
||||
\end{itemize}
|
||||
\item Global
|
||||
\begin{itemize}
|
||||
\item \# of points
|
||||
\item \# of strokes
|
||||
\item Center point
|
||||
\item Bitmap
|
||||
\item Bounding box (width, height, time)
|
||||
\end{itemize}
|
||||
\end{itemize}
|
||||
\end{frame}
|
||||
Binary file not shown.
|
After Width: | Height: | Size: 46 KiB |
|
|
@ -0,0 +1,30 @@
|
|||
\subsection{Online and offline recognition}
|
||||
|
||||
\begin{frame}{What is my Bachelor's thesis about?}
|
||||
\begin{itemize}
|
||||
\item Recognition of handwritten mathematical symbols
|
||||
\item On-line recognition, not OCR!
|
||||
\item Given a series of points $(x(t), y(t), b(t))$\\
|
||||
I want to get the proper \LaTeX{} command.
|
||||
\end{itemize}
|
||||
\end{frame}
|
||||
|
||||
\begin{frame}{Why did I work on this topic?}
|
||||
\begin{itemize}
|
||||
\item \LaTeX{} is easy as soon as you know the \textbackslash{}commands.
|
||||
\item It's hard to find the \LaTeX{} command of single symbols.
|
||||
\item It's much harder to find complete formulas.
|
||||
\end{itemize}
|
||||
|
||||
% I want to
|
||||
% \begin{itemize}
|
||||
% \item provide a tool that enables beginners to get the best \LaTeX{} command
|
||||
% for their formula,
|
||||
% \item find out what works best for symbol recognition
|
||||
% \item and provide data and a platform to test new ideas for classifiers
|
||||
% \end{itemize}
|
||||
|
||||
For now: recognition of isolated symbols. That means:
|
||||
|
||||
single symbol \enquote{formulas} rather than multi-symbol formulas
|
||||
\end{frame}
|
||||
Binary file not shown.
|
After Width: | Height: | Size: 46 KiB |
|
|
@ -0,0 +1,18 @@
|
|||
\subsection{Preprocessing}
|
||||
\begin{frame}{Preprocessing}
|
||||
\begin{itemize}
|
||||
\item Normalizing
|
||||
\begin{itemize}
|
||||
\item Scaling
|
||||
\item Shifting
|
||||
\item Resampling
|
||||
\end{itemize}
|
||||
\item Noise reduction
|
||||
\begin{itemize}
|
||||
\item Smoothing (e.g. moving average)
|
||||
\item Dot reduction
|
||||
\item Filtering (by distance, speed or angle)
|
||||
\item Stroke connection
|
||||
\end{itemize}
|
||||
\end{itemize}
|
||||
\end{frame}
|
||||
12
presentations/Bachelor-Final-Presentation/LaTeX/sources.tex
Normal file
12
presentations/Bachelor-Final-Presentation/LaTeX/sources.tex
Normal file
|
|
@ -0,0 +1,12 @@
|
|||
\subsection{Sources}
|
||||
\begin{frame}{Image Sources}
|
||||
\begin{itemize}
|
||||
\item \href{https://commons.wikimedia.org/wiki/File:Server-multiple.svg}{Server} by RRZEicons
|
||||
\item \href{https://commons.wikimedia.org/wiki/File:Computer-aj_aj_ashton_01.svg}{Desktop Computer} by Ed g2s,
|
||||
Ironbrother, Kierancassel and Msgj
|
||||
\item \href{https://commons.wikimedia.org/wiki/File:Server_by_mimooh.svg}{Server} by Mimooh
|
||||
\end{itemize}
|
||||
|
||||
The presentation can be found at \url{http://tinyurl.com/write-math-short-presentation}
|
||||
\end{frame}
|
||||
|
||||
24
presentations/Bachelor-Final-Presentation/LaTeX/will-do.tex
Normal file
24
presentations/Bachelor-Final-Presentation/LaTeX/will-do.tex
Normal file
|
|
@ -0,0 +1,24 @@
|
|||
\subsection{What will I do next?}
|
||||
\begin{frame}{What will I do next?}
|
||||
\begin{itemize}
|
||||
\item Include the currently best model in write-math.com
|
||||
\item Evaluate preprocessing steps
|
||||
\item Try other features
|
||||
\item Try other topologies / trainings (e.g. pretraining, newbob)
|
||||
\item Eventually try convolutional neural nets
|
||||
\end{itemize}
|
||||
\end{frame}
|
||||
|
||||
% \subsection{Far future}
|
||||
% \begin{frame}{What could be done?}
|
||||
% \begin{itemize}
|
||||
% \item Make use of audio data in a multimodal approach\\
|
||||
% e.g. $R$ and $\mathcal{R}$
|
||||
% \item Currently, the Lecture Translation system doesn't recognize math.\\
|
||||
% You get \enquote{integral of e raised to the power of x d x} instead
|
||||
% of $\int e^x \mathrm{d} x$.
|
||||
% \item Spoken math is ambigous: $\sqrt{a+b}$ vs. $\sqrt{a} + b$
|
||||
% \item The language model I create could help to find probable formulas
|
||||
% \item The platform could be used to get more input data of users
|
||||
% \end{itemize}
|
||||
% \end{frame}
|
||||
149
presentations/Bachelor-Final-Presentation/LaTeX/write-math.tex
Normal file
149
presentations/Bachelor-Final-Presentation/LaTeX/write-math.tex
Normal file
|
|
@ -0,0 +1,149 @@
|
|||
\subsection{Write Math}
|
||||
|
||||
\begin{frame}{write-math.com}
|
||||
\begin{itemize}
|
||||
\item a website where users can add labeled training data and unlabeled
|
||||
data which they want to classify. I call this data \enquote{recording}
|
||||
\begin{figure}[ht]
|
||||
\centering
|
||||
\subfloat{
|
||||
\includegraphics[height=0.1\textwidth]{../images/279952.pdf}
|
||||
}%
|
||||
\qquad
|
||||
\subfloat{
|
||||
\includegraphics[height=0.1\textwidth]{../images/281507.pdf}
|
||||
}%
|
||||
\qquad
|
||||
\subfloat{
|
||||
\includegraphics[height=0.1\textwidth]{../images/287612.pdf}
|
||||
}%
|
||||
\qquad
|
||||
\subfloat{
|
||||
\includegraphics[height=0.1\textwidth]{../images/292175.pdf}
|
||||
}%
|
||||
\caption*{4 recordings}
|
||||
\end{figure}
|
||||
\item works with desktop computers and touch devices
|
||||
\item symbol recognition can be done by multiple classifiers
|
||||
\item users can contribute formulas as recordings and as \LaTeX{} answers
|
||||
for recordings
|
||||
\item users can vote for \LaTeX{} answers:
|
||||
\Large $\leq$, $\leqq$, $\leqslant$, \dots \normalsize
|
||||
\item user who entered the recording can accept one answer
|
||||
\end{itemize}
|
||||
\end{frame}
|
||||
|
||||
% \framedgraphic{Classify}{../images/classify.png}
|
||||
% \framedgraphic{Workflow}{../images/workflow.png}
|
||||
% \framedgraphic{User page}{../images/user-page.png}
|
||||
% \framedgraphic{Information about recordings}{../images/view.png}
|
||||
% \framedgraphic{Symbol page}{../images/symbol.png}
|
||||
% \framedgraphic{Training}{../images/train.png}
|
||||
% \framedgraphic{Ranking}{../images/ranking.png}
|
||||
|
||||
|
||||
\begin{frame}[fragile]{Statistics}
|
||||
\begin{itemize}
|
||||
\item 127 users with at least 5 recordings
|
||||
\item $\num{1111}$ symbols, but only $\num{369}$ used for experiments
|
||||
\item $\num{235831}$ recordings (e.g. $\num{3489}$ times \verb+\int+, but only 50 times \verb+X+)
|
||||
\end{itemize}
|
||||
\end{frame}
|
||||
|
||||
% \begin{frame}{First classification worker}
|
||||
% \begin{itemize}
|
||||
% \item preprocessing: Scale to fit into unit square while keeping the aspect
|
||||
% ratio
|
||||
% \item applies greedy time warping
|
||||
% \item compares a new recording with every recording
|
||||
% in the database
|
||||
% \item[$\Rightarrow$] Classification time is in $\mathcal{O}(\text{recordings})$,
|
||||
% but we rather would like $\mathcal{O}(\text{symbols})$
|
||||
% \item the current server / workflow can only handle about 4000 recordings
|
||||
% \item[$\Rightarrow$] Another way to classify is necessary
|
||||
% \end{itemize}
|
||||
% \end{frame}
|
||||
|
||||
\begin{frame}[fragile]{Handwriting Recognition Toolkit (hwrt)}
|
||||
\begin{itemize}
|
||||
\item Many preprocessing algorithms / features implemented
|
||||
\item Plug-in system for preprocessing algorithms / features
|
||||
\item Needs neural network toolkit
|
||||
\item Hosted at \url{https://github.com/MartinThoma/hwrt}
|
||||
\item Installable via \verb+pip+ (Python package installer):\\
|
||||
\verb+$ pip install hwrt+
|
||||
\end{itemize}
|
||||
\end{frame}
|
||||
|
||||
\begin{frame}[fragile]{hwrt preprocessing configuration file}
|
||||
\begin{verbatim}
|
||||
data-source: raw-datasets/2014-08-26-20-14-data-raw.pickle
|
||||
queue:
|
||||
- RemoveDuplicateTime: null
|
||||
- ScaleAndShift:
|
||||
- max_width: 1.0
|
||||
- max_height: 1.0
|
||||
- center: true
|
||||
- SpaceEvenlyPerStroke:
|
||||
- kind: linear
|
||||
- number: 20
|
||||
\end{verbatim}
|
||||
\end{frame}
|
||||
|
||||
\subsection{Preprocessing algorithms}
|
||||
\begin{frame}{Preprocessing algorithms}
|
||||
\begin{itemize}
|
||||
\item Normalizing
|
||||
\begin{itemize}
|
||||
\item Scaling
|
||||
\item Shifting
|
||||
\item Resampling
|
||||
\end{itemize}
|
||||
\item Noise reduction
|
||||
\begin{itemize}
|
||||
\item Smoothing (e.g. moving average)
|
||||
\item Dot reduction
|
||||
\item Filtering (by distance, speed or angle)
|
||||
\item Stroke connection
|
||||
\end{itemize}
|
||||
\end{itemize}
|
||||
\end{frame}
|
||||
|
||||
\begin{frame}[fragile]{hwrt feature configuration file}
|
||||
\begin{verbatim}
|
||||
data-source: preprocessed/baseline
|
||||
data-multiplication:
|
||||
- Multiply:
|
||||
- nr: 1
|
||||
features:
|
||||
- ConstantPointCoordinates:
|
||||
- strokes: 4
|
||||
- points_per_stroke: 20
|
||||
- fill_empty_with: 0
|
||||
- pen_down: false
|
||||
\end{verbatim}
|
||||
\end{frame}
|
||||
|
||||
\subsection{Features}
|
||||
\begin{frame}{Features}
|
||||
\begin{itemize}
|
||||
\item Local
|
||||
\begin{itemize}
|
||||
\item Coordinates
|
||||
\item Speed
|
||||
\item Binary pen pressure
|
||||
\item Direction
|
||||
\item Curvature
|
||||
\item Bitmap-environment
|
||||
\item Hat-Feature
|
||||
\end{itemize}
|
||||
\item Global
|
||||
\begin{itemize}
|
||||
\item \# of points
|
||||
\item \# of strokes
|
||||
\item Center point
|
||||
\item Bitmap
|
||||
\item Bounding box (width, height, time)
|
||||
\end{itemize}
|
||||
\end{itemize}
|
||||
\end{frame}
|
||||
Loading…
Add table
Add a link
Reference in a new issue