3 papers
cs.CR2021
A framework for comprehensible multi-modal detection of cyber threats
Jan Kohout, Čeněk Škarda, Kyrylo Shcherbin +2
Detection of malicious activities in corporate environments is a very complex task and much effort has been invested into research of its automation. However, vast majority of exis…
cs.LG2021
Cluster Representatives Selection in Non-Metric Spaces for Nearest Prototype Classification
Jaroslav Hlaváč, Martin Kopp, Jan Kohout
The nearest prototype classification is a less computationally intensive replacement for the -NN method, especially when large datasets are considered. In metric spaces, centroi…
cs.LG2021
Learning Explainable Representations of Malware Behavior
Paul Prasse, Jan Brabec, Jan Kohout +3
We address the problems of identifying malware in network telemetry logs and providing \emph{indicators of compromise} -- comprehensible explanations of behavioral patterns that id…