Scientists Convert Computer Viruses into Images
www.posiv.cz 47
BUSINESS
into
Using a mathematical method,
the researchers converted these threats
into visual forms and then fed them to
artificial intelligence. The AI has become
very successful at identifying whether
the software is “good” or malicious,
i.e., malware. In addition to the visually
appealing representation of these digital
threats, the proposed method improves
the accuracy of their detection and
provides new insights into their behaviour.
“We have developed a method that
can monitor the dynamic behaviour
of malware and convert it using fractal
geometry—a branch of mathematics that
deals with very complex shapes and their
visualisation—into a very attractive visual
form. The AI then processed the images
and learned to distinguish between
good and bad software. We provided
it with about 130,000 images in two
types of experiments, half of which were
goodware and half malware. We then
presented it with entirely unknown viruses
and asked it to evaluate them. It was able
to identify malware with up to 91 percent
accuracy and is continuously improving,”
explained Ivan Zelinka from FEI, who,
along with his colleagues, published
the results in the journal Mathematics and
Computers in Simulation.
The study opens new avenues in
malware research, showing that fractal
geometry can significantly improve their
visualisation and classification. “As the field
of cybersecurity evolves and new threats
continue to emerge, such interdisciplinary
methods will be crucial for maintaining
an advantage over these dangers.
Therefore, we continue our research,
focusing on static analyses in addition
to dynamic ones, which are quicker for
practical detection of dangerous viruses,”
Zelinka clarified.
The methodology developed based on
dynamic analysis also has numerous
benefits. It provides information on many
details of malware behaviour in real-time.
“The method, with the help of dynamic
analysis, is important for further research
so that experts can analyse and study
the virus post-mortem. Additionally, it has
opened up a range of other interesting
research questions,” added computer
scientist and cyberneticist Zelinka.
Fractal geometry and fractals are used in
many scientific fields and inspire many
artists. Although the Ostrava scientist
does not consider himself an artist, he
admits to a fondness for visual art and
fractals. “These are very beautiful shapes,
and I am pleased when very abstract
concepts, or in this case, digital behaviour
in cyberspace, can take on such a visual
form. In this case, it is not merely for
aesthetics, but we have also gained
an effective tool for further advancing
cybersecurity,” concluded Zelinka.
We provided it with about 130,000 images in two types
of experiments, half of which were goodware and half
malware. We then presented it with enrely unknown
viruses and asked it to evaluate them. It was able to
idenfy malware with up to 91 percent accuracy and is
connuously improving.
aby odborníci mohli ex post virus analyzo-
vat a zkoumat. Navíc se nám otevřela řada
dalších zajímavých odborných otázek,“
doplnil informatik a kybernetik Zelinka.
Fraktální geometrie a fraktály se využí-
vají v řadě vědeckých oborů, ale inspi-
ruje se jimi i řada výtvarníků. Ostravský
vědec se mezi umělce sice neřadí, ale
zálibu ve výtvarném umění a fraktálech
přiznává. „Jedná se o velmi krásné ob-
razce a já jsem rád, když tak abstraktní
pojmy, nebo v tomto případě digitální
chování v kyberprostoru, mohou získat
takovouto vizuální podobu. V tomto
případě to navíc není samoúčelné, ale
získali jsme i účinný nástroj pro další roz-
voj kyberbezpečnosti,“ uzavřel Zelinka.