My research group (https://www.ivanzelinka.eu/NAVY/) has been working on computer viruses and unconventional algorithms, where fractal geometry belongs, for many years. Since about 2020, we have actively collaborated with ESET, which has provided us with several million dynamic malware analysis samples for further analysis and experiments. In 2021, the idea was conceived to elevate the visualization of malware, which until then had been visualized in the form of only black and white maps, reminiscent of the grain of an old TV, to the level of fractal geometry.
By the end of 2021, the first images of malware visualization using fractal geometry were created, where we visualized several hundred malware families as fractal patterns. In 2022, our research continued by creating a special software called Fractal Visualizer, which was able to automatically convert dynamic malware analyses into fractal patterns, so that during 2022, when we had already generated a total of 120,000 malware and goodware images from ESET's multi-million sample dataset. At that time, we had already created the first prototype of our research paper (now ArXiv - Have You Ever Seen Malware?, https://doi.org/10.48550/arXiv.2212.02341), and we had been discussing our results with ESET experts all last year.
We deployed artificial intelligence in deep learning on these samples and managed to classify goodware from malware in this image plane with a decent success rate. We were especially considering that these were the first attempts of this kind ever. Throughout the research, we briefly presented and discussed some parts with our colleagues abroad who are prominent researchers and editors of various journals such as Swagatam Das (India), Siddhartha Bhattacharyya (India), Donald Davendra (Central Washington University USA), Eslam Amer (Egypt), Tarek Gaber (GB), Rui Miguel Silva (Portugal) or Dr Giacomo Innocenti (Universita degli Studi di Firenze, Via S. Marta 3,50139 Firenze, Italy) where we presented our results in the fusion of artificial intelligence, cybersecurity and these unique methods of visualization using fractal geometry during a week-long visit to local research group and PhD students.
During our research, one of Russia's most cited researchers for 2016, Nikolay V. Kuznetsov, joined our group and contributed to our research results. In the middle of the year, we also published our results at the National Cybersecurity Platform Meeting organized by the National Office for Cyber and Information Security of the Czech Republic (https://www.nukib.cz/). In mid-2022, we also uploaded the first version of the paper to the arXiv platform. In this version, we intentionally omitted one section discussing how to accurately compute fractal patterns. Since then, virtually all information about our research has been publicly available via the arXiv platform, making it vulnerable to idea theft. Unfortunately, This has happened to us twice in the field of swarm intelligence research. We are further developing our research in visualizing and classifying malware and goodware using fractal geometry. We are also trying to find a suitable journal to publish our interim results, which we have obtained over two years of intensive research and communication with ESET.
Unfortunately, we find that even refurbished journals do not seem to be ready for interdisciplinary research bringing new ideas. Thus, we are constantly looking for a suitable journal that is open to new ideas combining different areas of science and research.
If you are interested in our research, we would be very happy for any reasonable professional communication and collaboration. Again, thank you for reading this report, and we wish you a pleasant viewing of some selected fractal-malware samples, which are shown below or you can watch our YouTube video here: https://youtu.be/GPjaIkO9fzg
All names of our fractal malware figures are artistic like Batman etc., however original names are technical like 18_Win32.Doomer.E worm ...
NAVY lab team :)
Below is just sample of our 120 000 vizualizations of different malwares and goodwares. We prepare here link to download them, meanwhile, we can send you figure samples on demand...