The detailed comparison task of PAN 2012 consisted in a comparison
of given document pairs, with the expected output being the annotation of
similarities found between these documents.
-The submitted program has been run in a controlled environment
+The submitted program was running in a controlled environment
separately for each document pair, without the possibility of keeping any
-data between runs.
+cachd data between runs.
%In this section, we describe our approach in the detailed comparison
%task. The rest of this section is organized as follows: in the next
\subsection{Differences Against PAN 2010}
Our approach in this task
-is loosely based on the approach we have used in PAN 2010 \cite{Kasprzak2010}.
+is loosely based on the approach we used in PAN 2010 \cite{Kasprzak2010}.
The main difference is that instead of looking for similarities of
one type (for PAN 2010, we have used word 5-grams),
-we have developed a method of evaluating multiple types of similarities
+we developed a method of evaluating multiple types of similarities
(we call them {\it common features}) of different properties, such as
density and length.
-As a proof of concept, we have used two types of common features: word
+As a proof of concept, we used two types of common features: word
5-grams and stop-word 8-grams, the later being based on the method described in
\cite{stamatatos2011plagiarism}.
-In addition to the above, we have made several minor improvements to the
+In addition to the above, we made several minor improvements to the
algorithm, such as parameter tuning and improving the detections
merging in the post-processing stage.
are associated: the offset where the word begins, and the word length.
The offset where the word begins is not necessarily the first letter character
-of the word itself. We have discovered that in the training corpus,
+of the word itself. We discovered that in the training corpus,
some plagiarized passages were annotated including the preceding
-non-letter characters. We have used the following heuristics to add
+non-letter characters. We used the following heuristics to add
parts of the inter-word gap to the previous, or the next adjacent word:
\begin{itemize}
-\item when the inter-word gap contains interpunction (any of the dot,
+\item When the inter-word gap contains interpunction (any of the dot,
semicolon, colon, comma, exclamation mark, question mark, or quotes),
add the characters up to, and including the interpunction, to the previous
word, ignore the space character(s) after the interpunction, and add
the rest to tne next word.
-\item otherwise, when the inter-word gap contains newline, add the character
+\item Otherwise, when the inter-word gap contains newline, add the character
before the first newline to the previous word, ignore the first newline
character, and add the rest to the next word.
\item otherwise, ignore the inter-word gap characters altogether.
\end{itemize}
-When the detection program has been presented three different
+When the detection program was given three different
files instead of two (meaning the third one is machine-translated
-version of the second one), we have tokenized the translated document instead
-of the source one. We have make use of the line-by-line alignment of the
+version of the second one), we tokenized the translated document instead
+of the source one. We used the line-by-line alignment of the
source and machine-translated documents to transform the word offsets
and lengths in the translated document to the terms of the source document.
(the, of, and, a, in, to), or the possessive {\it'{}s}.
\end{itemize}
-We have represented each feature with the 32 highest-order bits of its
+We represented each feature with the 32 highest-order bits of its
MD5 digest. This is only a performance optimization, targeted for
larger systems. The number of features in a document pair is several orders
of magnitude lower than $2^{32}$, so the probability of hash function
\subsection{Common Features}
-For further processing, we have taken into account only the features
+For further processing, we took into account only the features
present both in source and suspicious document. For each such
-{\it common feature}, we have created the list of
+{\it common feature}, we created the list of
$(\makebox{offset}, \makebox{length})$ pairs for the source document,
and a similar list for the suspicious document. Note that a given feature
can occur multiple times both in source and suspicious document.
To detect a plagiarized passage, we need to find a set of common features,
which map to a dense-enough interval both in the source and suspicious
-document. In our previous work, we have presented the algorithm
+document. In our previous work, we described the algorithm
for discovering these {\it valid intervals} \cite{Kasprzak2009a}.
A similar approach is used also in \cite{stamatatos2011plagiarism}.
-Both of these algorithms use the features of a single type, which
+Both of these algorithms use features of a single type, which
allows to use the ordering of features as a measure of distance.
When we use features of different types, there is no natural ordering
above algorithms, that the last character of the previous feature
is before the last character of the current feature, is broken.
-We have modified the algorithm for computing valid intervals with
+We modified the algorithm for computing valid intervals with
multi-feature detection to use character offsets
-only instead of feature order numbers. We have used valid intervals
+only instead of feature order numbers. We used valid intervals
consisting of at least 4 common features, with the maximum allowed gap
inside the interval (characters not belonging to any common feature
of a given valid interval) set to 4000 characters.
\label{postprocessing}
In the postprocessing phase, we took the resulting valid intervals,
-and made attempt to further improve the results. We have firstly
+and made attempt to further improve the results. We firstly
removed overlaps: if both overlapping intervals were
shorter than 300 characters, we have removed both of them. Otherwise, we
kept the longer detection (longer in terms of length in the suspicious document).
-We have then joined the adjacent valid intervals into one detection,
-if at least one of the following criteria has been met:
+We then joined the adjacent valid intervals into one detection,
+if at least one of the following criteria were met:
\begin{itemize}
\item the gap between the intervals contained at least 4 common features,
and it contained at least one feature per 10,000
\subsection{Other Approaches Explored}
-There are several other approaches we have evaluated, but which were
+There are several other approaches we evaluated, but which were
omitted from our final submission for various reasons. We think mentioning
-them here is worthwhile nevertheless.
+them here is worthwhile nevertheless:
\subsubsection{Intrinsic Plagiarism Detection}
plagdet score using the official script in Python took between 120 and
180 seconds, as there is no parallelism in this script.
-When we have tried to use intrinsic plagiarism detection and language
+When we tried to use intrinsic plagiarism detection and language
detection, the first phase took about 12500 seconds. Thus omitting these
featurs clearly provided huge performance improvement.
https://www.fi.muni.cz/~kas/git / pan12-paper.git / commitdiff